The Effects of Dual-Tasking on Gait Dynamics in Older Adults with

Cognitive Impairment

Tess C. Hawkins

Bachelor of Exercise & Sport Science, The University of Sydney

Master of Exercise Physiology, The University of Sydney

A thesis submitted to fulfill of the requirements for the degree of Master of Applied Science

(Research)

Faculty of Health Sciences,

Discipline of Exercise and Sports Science,

The University of Sydney

2019

-i-

SUPERVISORS’S STATEMENT

This is to certify that the thesis entitled “The effects of dual-tasking on gait dynamics in

older adults with cognitive impairment” submitted by Tess C. Hawkins in fulfilment of

the requirements for the degree of Masters by Research is in a form ready for examination.

Professor Maria Fiatarone Singh

Discipline of Exercise & Sport Science

Faculty of Health Sciences

The University of Sydney

-ii-

STATEMENT OF ORIGINALITY

I, Tess C. Hawkins, hereby declare that the work contained within this thesis is my own and

has not been submitted to any other university or institution as a part or a whole requirement

for any higher degree. I certify that the intellectual content of this thesis is the product of my

own work and that all the assistance received in preparing this thesis and sources have been

acknowledged.

In addition, ethical approval from the University of Sydney Human Ethics Committee was

granted for the study presented in this thesis. Participants were required to read a participant

information document and informed consent was gained prior to data collection.

Tess C. Hawkins

Discipline of Exercise & Sport Science

Faculty of Health Sciences

The University of Sydney

-iii-

ACKNOWLEDGMENTS

I would like to extend my sincere thanks and appreciation to all my supervisors Professor

Maria Fiatarone Singh, Dr Yorgi Mavros, Dr Trinidad Valenzuela Arteaga and Professor

Jeffrey Hausdorff. This process has been a continuous learning curve, and I feel incredibly

humbled to have had experienced educators and researchers guiding me throughout,

particularly within the complex world of gait dynamics. Knowing that I had your support was

ever encouraging. You’ve collectively taught me to always raise the bar, and I will continue

to do that.

I would also like to thank the diligent researchers that were part of the Study of Mental And

Resistance Training (SMART) study for their commitment. Likewise, I would like to thank

the participants in the SMART study for their involvement and effort, without their

participation I would not have been able to write this thesis.

Thank you to my research group, a team of generous colleagues committed to improving the

health of others and celebrating every occasion in style. Thank you to my friends in K121 for

providing endless motivation and constantly challenging my opinions.

Most importantly, I would like to thank my family for their unwavering support and

unconditional love. Finally, I would like to thank my wife, Francesca, for being there every

step of the way, acting as my moral sounding board and providing endless reassurance. I

cannot express how none of this would have been possible without you.

-iv-

TABLE OF CONTENTS

SUPERVISORS’S STATEMENT .......................................................................................... i

STATEMENT OF ORIGINALITY ........................................................................................ ii

ACKNOWLEDGMENTS ..................................................................................................... iii

LIST OF TABLES ................................................................................................................. vi

LIST OF FIGURES .............................................................................................................. vii

ABBREVIATIONS ............................................................................................................. viii

ABSTRACT ............................................................................................................................ x

CHAPTER 1: INTRODUCTION ........................................................................................... 1

1.1 RATIONALE ................................................................................................................ 1

1.2 BACKGROUND .......................................................................................................... 2

1.3 FINDINGS FROM PREVIOUS STUDIES ................................................................. 5

1.4 OBJECTIVES ............................................................................................................... 7

1.6 REFERENCES ............................................................................................................. 9

CHAPTER 2: SYSTEMATIC REVIEW.............................................................................. 14

2.1 AUTHOR CONTRIBUTION STATEMENT ............................................................ 15

2.2 ABSTRACT ................................................................................................................ 16

2.3 INTRODUCTION ...................................................................................................... 18

2.4 METHODS ................................................................................................................. 19

2.5 RESULTS ................................................................................................................... 27

2.6 DISCUSSION ............................................................................................................. 39

2.7 STRENGTHS ............................................................................................................. 44

-v-

2.8 LIMITATIONS ........................................................................................................... 44

2.9 CONCLUSIONS......................................................................................................... 45

2.10 REFERENCES ......................................................................................................... 47

CHAPTER 3: STUDY .......................................................................................................... 94

3.1 AUTHOR CONTRIBUTION STATEMENT ............................................................ 95

3.2 PREAMBLE ............................................................................................................... 96

3.3 ABSTRACT ................................................................................................................ 97

3.4 INTRODUCTION ...................................................................................................... 99

3.5 METHODS ................................................................................................................. 99

3.6 RESULTS ................................................................................................................. 107

3.7 DISCUSSION ........................................................................................................... 109

3.8 CONCLUSIONS AND IMPLICATIONS ................................................................ 114

3.9 REFERENCES ......................................................................................................... 115

CHAPTER 4: CONCLUSIONS AND IMPLICATIONS .................................................. 129

4.1 CONCLUSIONS....................................................................................................... 129

4.2 IMPLICATIONS AND FUTURE DIRECTIONS ................................................... 131

4.3 REFERENCES ......................................................................................................... 134

BIBLIOGRAPHY ........................................................................................................... 136

-vi-

LIST OF TABLES

Table 2.1 Medline full electronic search strategy example ........................... 61

Table 2.2 Cohort characteristics ..................................................................... 62

Table 2.3 Risk of bias assessment .................................................................. 66

Table 2.4 Dual-task procedure characteristics ................................................ 68

Table 2.5a Gait outcomes for Mild Cognitive Impairment group: single-task

vs. dual-task comparisons ............................................................... 72

Table 2.5b Gait outcomes for dementia (including Alzheimer’s disease)

group: single-task vs. dual-task comparisons ................................. 76

Table 2.6a Gait outcomes for cognitive status comparisons: Control vs. Mild

Cognitive Impairment ..................................................................... 84

Table 2.6b Gait outcomes for cognitive status comparisons: Control vs.

dementia (including Alzheimer’s disease)...................................... 87

Table 2.7 Gait outcome: within study MCI vs. dementia ............................... 90

Table 2.8 Stratification of outcomes for meta-analysis .................................. 92

Table 3.1 Participant characteristics ............................................................. 126

Table 3.2 Factors significantly associated with changes in at least one

measure of gait variability and dynamics during dual-tasking ..... 127

-vii-

LIST OF FIGURES

Figure 2.1 Flow diagram of the systematic review process ............................. 56

Figure 2.2 Forest plot for within study comparison for single-task and dual-

task for all gait dynamic outcomes and all dual-task procedures .. 57

Figure 2.3 Forest plot for within study comparison for cognitively impaired

and cognitively healthy for all gait dynamic outcomes and all

dual-task procedures ....................................................................... 59

Figure 2.4 Forest plot for meta-analysis of within study comparison for MCI

and dementia for all gait dynamic outcomes and all dual-task

conditions ........................................................................................ 60

Figure 3.1 Gait dynamics and cognitive performance under single-task and

dual-task walking conditions ........................................................ 122

Figure 3.2 Relationship between gait dynamics and brain morphology ........ 124

-viii-

ABBREVIATIONS

a-MCI Amnestic Mild Cognitive Impairment

AD Alzheimer’s disease

AKA Above knee amputation

BKA Below knee amputation

CI Confidence interval

CMT Charcot-Marie-Tooth

CV Coefficient of variations

DFA Detrended fluctuation analysis

DSM-IV Diagnostic and Statistical Manual of Mental Disorders (4th edition)

EF Executive function

EPS Extra-pyramidal signs

ES Effect size

FTD Frontotemporal dementia

GDS Geriatric Depression Scale

HADS Hospital Anxiety and Depression Scale

IEF Impaired executive function

MCI Mild Cognitive Impairment

MD Mean difference

MMSE Mini-mental State Exam

MNA Mini Nutritional Assessment

MS Multiple sclerosis

Na-MCI Non-amnestic Mild Cognitive Impairment

-ix-

NINCDS-ADRDA National Institute of Neurological and Communicative Disorders and

Stroke and the Alzheimer's Disease and Related Disorders

Association

SMD Standardized mean difference

SD Standard deviation

TUG Timed Up and Go

-x-

ABSTRACT

Impaired gait dynamics are associated with increased falls risk, and are worse in cognitively

impaired older adults. Dual-tasking is the performance of a second task, cognitive or physical,

while walking. Dual-tasking impairs gait in individuals with deficits in cognitive function,

and may reveal abnormalities in gait dynamics not observed under single-task conditions,

known as ‘dual-task cost’.

The aims of this thesis were to review the literature on dual-task costs on gait dynamics in

adults with cognitive impairment, and to identify clinical characteristics associated with this

cost in older adults with Mild Cognitive Impairment (MCI) or dementia.

First, a systematic review of 25 articles that measured single- and dual-task walking in adults

with MCI or dementia was conducted. Findings suggested that gait is worsened under dual-

task conditions compared to single-task conditions. Furthermore, dual-task cost is higher in

individuals with cognitive impairment compared to cognitively healthy older adults, and

similarly, higher in adults with dementia compared to adults with MCI. Research is lacking

into nonlinear gait dynamics, the relationship to fall risk, and other characteristics which may

be associated dual-task gait dynamics. Next, data from an acute exposure to dual-tasking in

93 adults with MCI were used to explore linear and nonlinear effects of dual tasking on gait

dynamics. Gait dynamics were assessed using stride time variability and detrended

fluctuation analyses fractal scaling exponent. Cognitive, physical and psychosocial function

and brain morphology were assessed to identify any associations with gait dynamics. Gait

dynamics worsened significantly during dual-tasking, while cognitive performance was

preserved. Additionally, a higher dual-task cost of gait dynamics was associated with lower

-xi-

aerobic fitness, poorer balance, reduced psychological well-being, and reductions in brain

thickness and volume in the posterior cingulate and hippocampus respectively.

Dual-task costs are accentuated in the presence of cognitive dysfunction. Observed

associations with physical fitness, psychological well-being and brain volumes suggest that

interventions targeting these modifiable characteristics could potentially improve dual-task

performance, and ultimately falls risk, in adults with cognitive impairment.

-1-

CHAPTER 1: INTRODUCTION

1.1 RATIONALE

One-third of adults over 65 years of age fall each year [1, 2], and that fall risk is doubled

among older adults with cognitive impairment [3]. Approximately 10% of falls are injurious

[4], and as a leading cause of injury-related hospitalisations [5], the overall cost of falls in the

United States of America is approximately US$50 billion [6], and more than $600 million

per year in Australia [7].

The high global prevalence of falls has prompted research into the prevention and

identification of risk factors associated with falls [8]. Potential risk factors for older adults

include, but are not limited to, sarcopenia, dizziness, gait dysfunction, visual disorders,

postural hypotension, balance impairment and cognitive dysfunction [9]. The multifactorial

nature of falls has been well studied in cognitively healthy older adults, however, less is

known about the nature and interrelationship of risk factors in cognitively impaired older

adults [10].

Variations in gait are important because they contribute to mobility and functional

impairment, and predispose individuals to falls [11]. When a secondary task is added to

walking, cortical control may be challenged and walking regulation may become worse,

leading to a further increase in risk in those already predisposed to fall [12]. Additionally,

changes in cognitive functioning contribute to an increased risk of falls, with risk of falls

heightened under dual-task walking conditions [13]. Dual-task walking is the performance of

a second, concurrent task while simultaneously walking. Therefore, people with cognitive

impairment are more likely to have a gait pattern that is variable, especially under dual-task

-2-

conditions, which may increase their risk of falls. Therefore, this thesis specifically focuses

on gait characteristics in relation to walking under single and dual-task conditions and the

relationships between cognitive impairment and falls risk, as well as demographic,

psychological and physiological factors.

1.2 BACKGROUND

Gait dynamics

A healthy gait appears to flow effortlessly and rhythmically and is characterized by an upright

posture and freely swinging legs [14]. The analysis of gait is commonly performed in clinical

settings to assist in the risk assessment of falls and other neuro-motor outcomes [15].

Measurement methods can vary with assessment choice dependent on the purpose of the

assessment, the cost and usability of equipment, and the assessment environment.

Independent of gait assessment, gait characteristics are usually defined as either spatial,

temporal and, further, as linear nonlinear [16]. Examples of spatial aspects include step length

and step width, while temporal aspects include stride time and swing time. Nonlinear

measures include the unique analysis of temporal gait aspects to determine stability using

fluctuation metrics such as the calculation of a fractal scaling exponent using detrended

fluctuation analyses (DFA) of stride time.

One purpose of gait analysis is to better understand the dynamics of an individual’s gait. Gait

dynamics is a broad term used to describe the magnitude of stride-to-stride fluctuations and

their change over time during walking [4]. It encompasses all aspects of gait variability and

nonlinear changes. Gait variability is identified by intra-individual stride-to-stride

fluctuations in walking parameters such as stride time and step length [17]. Gait variability

assumes larger stride-to-stride fluctuations reflect poorer control of gait. Variations can be

observed during different aspects of the gait cycle and impact the dynamics of gait. Gait

-3-

variability outcomes are commonly measured by distributional metrics, including coefficient

of variation and standard deviation [18]. Distributional metrics are used for linear systems

and quantify the magnitude in variation in a set of spatial or temporal outcomes independently

of the order in the distribution [19]. Alternatively, nonlinear measures quantify the degree of

randomness in highly non-stationary physiological data. Stride interval time series data

exhibit long-range, power law correlations in the gait rhythm [20]. The self-similar, or fractal,

correlations highlight the presence of a ‘memory’ in the neurophysiological locomotor

control system where fluctuations are related to variations in the stride interval hundreds of

strides earlier [4]. Specific tools including DFA, approximate entropy and largest Lyapunov

exponent, have been developed for nonlinear systems to determine variation in how a motor

behavior emerges in time [19, 21]. Nonlinear measures aim to eliminate temporal trends,

which avoids the detection of correlations from non-stationary artifacts.

Recognition of the complex and multifactorial nature of gait dynamics has led to many

investigations to understand why such stride-to-stride or step-to-step variations occur and

what clinical implications they have [17]. Linear and nonlinear measures of gait dynamics

use different methods to assess and quantify the changes in gait variability and stability. As

described above, linear measures capture the magnitude of the changes and nonlinear

measures capture changes over time. The collection of both linear and nonlinear types of gait

outcomes are necessary to completely understand gait dynamics and associated factors.

Cognitive impairment

There is evidence that a more variable gait pattern is associated with cognitive decline in

older adults, particularly with the executive function domain [22]. Executive function (EF)

refers to the cognitive skills responsible for the planning and sequencing goal-oriented tasks

-4-

and the execution of complex activities. Executive function is most often operationalized by

assessments of working memory, attentional controls, and response inhibition [23]. Executive

function impairment has been associated with poorer performance in both spatial and

temporal gait variability measures in community-dwelling older adults [24], as well as in

individuals with the diagnosis of Mild Cognitive Impairment (MCI), dementia and

Alzheimer’s disease (AD) [25]. Mild cognitive impairment reflects a transitional state

between normal aging and AD [26], as well as being a prodrome to other forms of dementia,

including vascular dementia, particularly in subcortical microvascular disease [27].

According to Peterson (1999) [26], MCI can be defined as the following: (1) self-reported

memory complaints, preferably verified by an informant, (2) objective memory disorder, (3)

absence of other cognitive disorders and normal function in everyday life, (4) normal general

cognitive function and (5) absence of dementia. Individuals with MCI experience declines in

cognitive function of 6–10% per year compared to 1–2% in cognitively healthy older adults

[28], resulting in a 10-15 times higher risk of developing AD [29].

The sequencing of cognitive and gait impairments is not clear in MCI as most studies to date

have been cross-sectional, and thus it is not known whether executive function decline

precedes or follows changes in gait stability. Executive function contributes to the

performance of normal walking, and works collectively with sensorimotor systems and other

cognitive domains, i.e., attention, to ensure safe and efficient gait [30]. It is possible that

impaired gait dynamics may be an early sign of brain pathology which precedes the

manifestation of overt cognitive difficulties, and it is thus important to comprehend the

relationship of these two domains. Additionally, given the higher risk of falling in individuals

with cognitive impairment [31], it is important to understand in this cohort specifically,

-5-

whether the performance of complex tasks, involving simultaneous targeting of different

aspects of motor control and cognition, increase fall risk, as reviewed below.

Dual-task walking

Lundin-Olsson and colleagues [2] first reported that nursing home residents who were

observed clinically to stop talking when walking were at a significantly higher risk of falls

over the next 6 months, compared to those subjects who were able to walk and talk. Since

then, dual-tasking, including ‘walking while talking’ has been investigated as a potential

marker of mobility, cognitive impairment and fall risk [32]. Dual-tasking, the performance

of two tasks simultaneously, is a clinically relevant condition that attempts to recreate in the

clinic or laboratory situations where community-dwelling older adults are at a heightened risk

of falling [17]. Dual-tasking has been shown to increase gait variability in adults, especially

in individuals with cognitive impairment. For example, temporal gait variability measures are

significantly worsened under dual task conditions in older adults with MCI and AD compared

to age-matched normal controls [33]. Dual tasking impairments are a known motor

characteristic of MCI, and are a potential marker for further cognitive decline [34].

1.3 FINDINGS FROM PREVIOUS STUDIES

Gait dynamics have been shown to remain relatively stable over time in cognitively healthy

individuals [35] but to worsen over time in cognitively impaired individuals [36].

Additionally, gait dynamics [37] and risk of falls are increased in older adults with MCI

compared to cognitively healthy older adults [38], with those who are cognitively impaired

and have fallen having worse gait dynamics compared to those who are cognitively impaired

but have not fallen [39]. Gait changes under dual-task conditions have also been associated

with future fall risk in cognitively healthy community-dwelling older adults [40]. Linear and

-6-

nonlinear measures of gait dynamics may hold the predictive ability to distinguish between

healthy and fall-prone older adults [16].

Stride time variability [41] and stride length variability [42] have been repeatedly shown to

increase under dual-task conditions compared to single-task conditions. Such dual-task

conditions increase gait variability for both healthy adults and adults with MCI, however,

decrements in performance are different across the cognitive spectrum, with larger

decrements observed in those who are cognitively impaired [33, 37, 43]. Notably, under

dual-task conditions cognitively healthy older adults have been reported to prioritize

cognition over motor performance, referred to as the ‘cognitive-first’ approach [44]. This

pattern is also seen in adults with MCI, who when asked to prioritize either their gait or

cognitive performance during a dual-task experiment, increase gait variability just as they do

with dual-tasking with no instructions regarding task prioritisation [43].

Recent reviews of the literature have focused on the relationships between gait dynamics and

individuals with neurological disorders [18], and use of interventions to target dual-task

performance [45-47]. Several studies implementing interventions targeting dual-task

performance have shown improvements in dual-task gait dynamics in healthy individuals [46]

and individuals with neurodegenerative disease [47]. However, there are too few studies to

generalize the effects of the interventions. Muir-Hunter and colleagues [40] noted a lack of

evidence-based recommendations for dual-task testing to evaluate fall risk in clinical practice.

Finally, research has begun to explore gait characteristics associated with falls in dementia,

including AD [48]. For example, step length variability has been associated with recurrent

falls and the use of mobility aids and walking outdoors, and reduced walking frequency

(amount of time spent walking) has been associated with increased falls risk in individuals

-7-

with dementia [48]. There is great interest in preventing falls by studying a cohort at high risk

for both dementia and gait dynamics under dual-task conditions: older adults with MCI.

Understanding the clinical characteristics which are associated with dual-task gait dynamics

in older adults with MCI is critical to the development of preventative strategies that will

hopefully preserve gait despite intrinsic and environmental stressors known to trigger falls.

1.4 OBJECTIVES

To prevent falls and improve gait dynamics in those at risk, an understanding of the changes

which occur in people with cognitive impairment when dual-tasking is necessary to identify

potential targets for the development of better fall reduction strategies. The aim of this thesis

was to advance the knowledge of dual-task gait dynamics in older adults with mild cognitive

impairment and dementia, and to identify clinical and physiological characteristics associated

with these dynamics.

The following objectives were investigated within this thesis:

1. To evaluate the effect of dual-task walking on changes in gait dynamics, termed dual-

task cost, for older adults with cognitive impairment;

2. To explore if the degree of cognitive impairment, dual-task paradigm and/or gait

dynamic measure influences the dual-task cost of gait dynamics in cognitively

impaired older adults;

3. To identify characteristics associated with the dual-task gait dynamics of cognitively

impaired older adults;

4. To determine the extent of changes in gait and the performance of a secondary

cognitive task during dual-task walking in older adults with cognitive impairment;

-8-

5. To explore associations with dual-task performance/costs, specifically strength,

aerobic capacity, functional performance, psychosocial function, and brain

morphology in adults with cognitive impairment.

-9-

1.6 REFERENCES

1. Blake, A., et al., Falls by elderly people at home: prevalence and associated factors.

Age and Ageing, 1988. 17(6): p. 365-372.

2. Lundin-Olsson, L., L. Nyberg, and Y. Gustafson, “Stops walking when talking” as a

predictor of falls in elderly people. The lancet, 1997. 349(9052): p. 617.

3. Shaw, F.E., Falls in cognitive impairment and dementia. Clinics in Geriatric

Medicine, 2002. 18(2): p. 159-173.

4. Hausdorff, J.M., Gait dynamics, fractals and falls: Finding meaning in the stride-to-

stride fluctuations of human walking. Human Movement Science, 2007. 26(4): p. 555-

589.

5. Pointer, S.C., Trends in hospitalised injury, Australia 1999-00 to 2014-15, in Injury

research and statistics series no. 110. 2018, AIHW: Canberra.

6. Florence, C.S., et al., Medical Costs of Fatal and Nonfatal Falls in Older Adults.

Journal of the American Geriatrics Society, 2018. 66(4): p. 693-698.

7. Bradley, C. and J. Harrison, Hospitalisations due to falls in older people, Australia,

2003-2004, in Injury research and statistics series no. 32. 2007, AIHW: Canberra.

8. Stewart Williams, J., et al., Prevalence, risk factors and disability associated with

fall-related injury in older adults in low- and middle-incomecountries: results from

the WHO Study on global AGEing and adult health (SAGE). BMC Medicine, 2015.

13(1): p. 147.

9. Rubenstein, L.Z., Falls in older people: epidemiology, risk factors and strategies for

prevention. Age and Ageing, 2006. 35(S2): p. 37-41.

10. Montero-Odasso, M., et al., Gait and cognition: A complementary approach to

understanding brain function and the risk of falling. Journal of the American

Geriatrics Society, 2012. 60(11): p. 2127-2136.

11. Hausdorff, J.M., D.A. Rios, and H.K. Edelberg, Gait variability and fall risk in

community-living older adults: A 1-year prospective study. Archives of Physical

Medicine and Rehabilitation, 2001. 82(8): p. 1050-1056.

-10-

12. Montero-Odasso, M., et al., Dual-tasking and gait in people with mild cognitive

impairment. The effect of working memory. BMC Geriatrics, 2009. 9(41).

13. Lamoth, C.J., et al., Gait stability and variability measures show effects of impaired

cognition and dual tasking in frail people. Journal of Neuroengineering &

Rehabilitation, 2011. 8: p. 2.

14. Pirker, W. and R. Katzenschlager, Gait disorders in adults and the elderly. Wiener

Klinische Wochenschrift, 2017. 129(3): p. 81-95.

15. Maki, B.E., Gait Changes in Older Adults: Predictors of Falls or Indicators of Fear?

Journal of the American Geriatrics Society, 1997. 45(3): p. 313-320.

16. Hamacher, D., et al., Kinematic measures for assessing gait stability in elderly

individuals: a systematic review. Journal of the Royal Society Interface, 2011. 8(65):

p. 1682-1698.

17. Hausdorff, J.M., Gait variability: methods, modeling and meaning. Journal of

NeuroEngineering and Rehabilitation, 2005. 2(1): p. 19.

18. Moon, Y., et al., Gait variability in people with neurological disorders: A systematic

review and meta-analysis. Human Movement Science, 2016. 47: p. 197-208.

19. Stergiou, N. and L.M. Decker, Human movement variability, nonlinear dynamics, and

pathology: Is there a connection? Human Movement Science, 2011. 30(5): p. 869-

888.

20. Hausdorff, J.M., et al., When human walking becomes random walking: fractal

analysis and modeling of gait rhythm fluctuations. Physica A: Statistical Mechanics

and its Applications, 2001. 302(1): p. 138-147.

21. Hausdorff, J.M., et al., Fractal dynamics of human gait: stability of long-range

correlations in stride interval fluctuations. Journal of Applied Physiology, 1996.

80(5): p. 1448-1457.

22. Byun, S., et al., Gait Variability Can Predict the Risk of Cognitive Decline in

Cognitively Normal Older People. Dementia and Geriatric Cognitive Disorders,

2018: p. 251-261.

-11-

23. Royall, D.R., et al., Executive Control Function. The Journal of Neuropsychiatry and

Clinical Neurosciences, 2002. 14(4): p. 377-405.

24. Martin, K.L., et al., Cognitive Function, Gait, and Gait Variability in Older People:

A Population-Based Study. The Journals of Gerontology: Series A, 2013. 68(6): p.

726-732.

25. Yogev-Seligmann, G., J.M. Hausdorff, and N. Giladi, The role of executive function

and attention in gait. Movement Disorders, 2008. 23(3): p. 329-342.

26. Petersen, R.C., Mild Cognitive Impairment: Clinical Characterization and Outcome.

The Journal of the American Medical Association, 1999. 281(19): p. 303-308.

27. Meyer, J.S., et al., Is Mild Cognitive Impairment Prodromal for Vascular Dementia

Like Alzheimer's Disease? Stroke, 2002. 33(8): p. 1981-1985.

28. Lehrner, J., et al., Annual conversion to Alzheimer disease among patients with

memory complaints attending an outpatient memory clinic: the influence of amnestic

mild cognitive impairment and the predictive value of neuropsychological testing.

Wiener Klinische Wochenschrift, 2005. 117(18): p. 629-635.

29. Petersen, R.C., et al., Current concepts in mild cognitive impairment. Archives of

Neurology, 2001. 58(12): p. 1985-1992.

30. Snijders, A.H., et al., Neurological gait disorders in elderly people: clinical approach

and classification. The lancet neurology, 2007. 6(1): p. 63-74.

31. Muir, S.W., K. Gopaul, and M.M. Montero Odasso, The role of cognitive impairment

in fall risk among older adults: a systematic review and meta-analysis. Age and

Ageing, 2012. 41(3): p. 299-308.

32. Ayers, E.I., et al., Walking while Talking and Falls in Aging. Gerontology, 2014.

60(2): p. 108-113.

33. Muir, S.W., et al., Gait assessment in mild cognitive impairment and Alzheimer's

disease: The effect of dual-task challenges across the cognitive spectrum. Gait and

Posture., 2012.

-12-

34. Montero-Odasso, M., et al., The motor signature of mild cognitive impairment: results

from the gait and brain study. Journals of Gerontology Series A-Biological Sciences

& Medical Sciences, 2014. 69(11): p. 1415-1421.

35. Herssens, N., et al., Do spatiotemporal parameters and gait variability differ across

the lifespan of healthy adults? A systematic review. Gait & Posture, 2018. 64: p. 181-

190.

36. van Iersel, M.B., et al., Systematic review of quantitativeclinical gait analysis in

patients with dementia. Zeitschrift fur Gerontologie und Geriatrie, 2004. 37(1): p. 27-

32.

37. Montero-Odasso, M., S.W. Muir, and M. Speechley, Dual-task complexity affects gait

in people with mild cognitive impairment: the interplay between gait variability, dual

tasking, and risk of falls. Archives of Physical Medicine & Rehabilitation, 2012.

93(2): p. 293-299.

38. Liu-Ambrose, T.Y., et al., Increased risk of falling in older community-dwelling

women with mild cognitive impairment. Physical Therapy, 2008. 88(12): p. 1482-

1491.

39. Allali, G., et al., Falls, Cognitive Impairment, and Gait Performance: Results From

the GOOD Initiative. Journal of the American Medical Directors Association, 2017.

18(4): p. 335-340.

40. Muir-Hunter, S.W. and J.E. Wittwer, Dual-task testing to predict falls in community-

dwelling older adults: a systematic review. Physiotherapy, 2016. 102(1): p. 29-40.

41. Al-Yahya, E., et al., Cognitive motor interference while walking: A systematic review

and meta-analysis. Neuroscience and Biobehavioral Reviews, 2011. 35(3): p. 715-

728.

42. Smith, E., et al., The Influence of a Cognitive Dual Task on the Gait Parameters of

Healthy Older Adults: A Systematic Review and Meta-Analysis. Journal of Aging &

Physical Activity, 2017. 25(4): p. 671-686.

-13-

43. Lee, J. and S. Park, Effects of a priority-based dual task on gait velocity and

variability in older adults with mild cognitive impairment. Journal of Exercise

Rehabilitation, 2018. 14(6): p. 993-997.

44. Corp, D.T., et al., Reduced motor cortex inhibition and a ‘cognitive-first’

prioritisation strategy for older adults during dual-tasking. Experimental

Gerontology, 2018. 113: p. 95-105.

45. Booth, V., V. Hood, and F. Kearney, Interventions incorporating physical and

cognitive elements to reduce falls risk in cognitively impaired older adults: A

systematic review. JBI Database of Systematic Reviews and Implementation Reports,

2016. 14(5): p. 110-135.

46. Plummer, P., et al., Effects of Physical Exercise Interventions on Gait-Related Dual-

Task Interference in Older Adults: A Systematic Review and Meta-Analysis.

Gerontology, 2015. 62(1): p. 94-117.

47. Wajda, D.A., et al., Intervention modalities for targeting cognitive-motor interference

in individuals with neurodegenerative disease: a systematic review. Expert Review

of Neurotherapeutics, 2017. 17(3): p. 251-261.

48. Modarresi, S., et al., Gait parameters and characteristics associated with increased

risk of falls in people with dementia: a systematic review. International

Psychogeriatrics, 2018: p. 1-17.

-14-

CHAPTER 2: SYSTEMATIC REVIEW

The Effects of Dual-Task Walking on Gait Dynamics in Older Adults with Cognitive

Impairment: A Systematic Review

Authors and contributions:

Tess C Hawkins MExPhys, Maria A Fiatarone Singh MD, Trinidad Valenzuela Arteaga PhD,

Jeffrey M Hausdorff PhD, and Yorgi Mavros PhD.

TH, YM, and MFS contributed to the conception of the work. All authors contributed to the

acquisition, analysis, or interpretation of data. TH drafted the manuscript and all authors

critically revised the manuscript. All authors gave final approval and agree to be accountable for

all aspects of work ensuring integrity and accuracy.

-15-

Faculty of Health Sciences

2.1 AUTHOR CONTRIBUTION STATEMENT

Candidate Name: Tess C Hawkins

Degree Title: Master of Applied Science (Research)

Paper Title: The effects of dual-task walking on gait dynamics in older adults with cognitive

impairment: A systematic review

As the research supervisor of the above candidate, I confirm that the above candidate has made

the following contributions to the above paper title:

- Conception and design of the research

- Analysis and interpretation of the findings

- Writing the paper and critical appraisal of content

Professor Maria Fiatarone Singh

Discipline of Exercise & Sport Science

Faculty of Health Sciences

The University of Sydney

-16-

2.2 ABSTRACT

Objectives

Cognitively impaired individuals have greater variability in gait dynamics than cognitively

healthy individuals. The stress of dual-tasking reveals abnormalities in gait dynamics not

observed under single-task conditions, and may be particularly relevant in individuals with

cognitive impairment. We aimed to review the cost of dual-tasking on gait dynamics in

cognitively impaired older adults, according to cognitive function diagnosis, type of dual-task

paradigm used and/or method of gait dynamics measurement, as well as to identify any clinical

characteristics associated with dual-task costs.

Method

This systematic review was conducted according to the PRISMA guidelines and prospectively

registered in the PROSPERO database (no. CRD42018105787). An electronic database search

was conducted on the 9th August 2018 in the following databases; Ageline, CINAHL, EBM

review database CCRCTs, EMBASE, Medline, PEDro, PsychINFO, Scopus and Web of

Science. An email alert system for new published articles was set up, with the last record from

this alert system screened on the 15th February 2019. A random-effects meta-analysis was

performed when I2 was <75%, alternatively, a narrative synthesis was undertaken.

Results

Among 16,519 citations, 25 articles met the inclusion criteria. All studies included a single-

and dual-task walking condition for a cognitively impaired group [Mild Cognitive Impairment

(MCI) or dementia]. Fourteen studies included a cognitively healthy comparison group and

-17-

3studies included both an MCI group and a dementia group. Twenty-seven different

spatiotemporal and nonlinear measures of gait dynamics and 20 different dual-task procedures

were identified in this literature. Gait dynamics in cognitively impaired older adults are

worsened under dual-task conditions compared to single-task conditions. In individuals with

cognitive impairment, the dual-task cost is higher than it is in healthy older adults. A meta-

analysis included the 3 studies that allowed for a direct comparison between adults with

dementia and MCI, and showed a significantly greater dual-task gait cost in dementia vs. MCI,

with a relative effect size of 0.60 [0.22, 0.99].

Conclusions

Cognitive disease severity increases as gait dynamics become more impaired. Research is

lacking into nonlinear gait dynamics, the relationship to fall risk, and other characteristics which

may be associated dual-task gait dynamics. More well-designed longitudinal studies and

controlled trials with adequately powered samples are needed to confirm the clinical utility and

predictive value of dual-task gait testing, as well as to provide a consensus on the most robust

methods of dual-tasking and gait outcome assessment.

Key words

Dual-task, Gait variability, Cognitive impairment, MCI, Dementia.

-18-

2.3 INTRODUCTION

Consistent, safe walking is essential for older adults to maintain independent living and avoid

falls [1]. Falls are a global public health concern, due to the increasing number of older people

world-wide [2]. One-third of adults over 65 years of age fall each year [3], resulting in an annual,

overall cost of more than $600 million in Australia [4]. Although falls are multifactorial, stride-

to-stride fluctuations within an individual’s walking pattern (also referred to as gait dynamics)

[5] reduce the stability of gait and the ability to resist perturbations or stressors leading to falls.

One such stressor is dual-tasking, which refers to the performance of two tasks simultaneously,

such as walking and talking. Dual-task paradigms are used in research to recreate situations

where adults are at increased risk of falling [6]. In particular, dual-task walking has been

investigated as a potential marker of mobility and falls risk in individuals with cognitive

impairment [7], where the additional task acts as a cognitive stressor, diverting attention away

from stable locomotion in those who have fewer cognitive reserves to cope with such stressors

[8]. Gait dynamics are worsened under these dual-task conditions, with future fall risk predicted

by the magnitude of gait instability induced by dual-tasking [9].

Thus, the gait dynamic response to dual-tasking has been used to explore the relationships

between motor and cognitive function. For example, dual-task gait dynamics are worse in

cognitively impaired older adults than in their healthy counterparts [10, 11]. This dual-task

deficit, called ‘dual-task cost’, has been identified as a motor characteristic of cognitive

impairment, specifically mild cognitive impairment (MCI) [8]. Mild cognitive impairment

reflects a transitional state between normal aging and AD [12], and can be further described as

amnestic or non-amnestic [13]. Given the doubled risk of falling in adults with cognitive

-19-

impairment compared to healthy adults [14, 15], it is important to understand whether

performance of complex tasks, involving simultaneous targeting of different aspects of motor

control and cognition, is a major mediator of this increased fall risk. Despite the reported

associations [10] between dual-task cost and falls risk in adults with cognitive impairment, there

are limitations and inconsistencies within the literature. Previous reviews have summarized the

effects of dual-tasking on gait dynamics in neurological populations [16-18], attempted to

determine if dual-tasking can help discriminate adults with MCI from cognitively healthy adults

[19] and adults with dementia [20], as well as the association of dual-task walking on risk of on

falls [21]. However, no review to our knowledge has solely focused on the effects of dual-tasking

on linear and nonlinear gait dynamics in cognitively impaired populations free from other

neurological diseases.

Therefore, the purpose of this review was to investigate the effects of dual-task walking on gait

dynamics in cognitively impaired adults, with detailed assessment of the different gait dynamics

measures, dual-task conditions and levels of cognitive impairment. Our objectives were: 1) to

evaluate dual-task costs in older adults with cognitive impairment, 2) to investigate differences

in cost attributable to cognitive function diagnosis, type of dual-task paradigm used and/or

method of gait dynamics measurement, and 3) to determine physiological and clinical

characteristics related to dual-task cost in cognitively impaired older adults. To our knowledge

this is the first systematic review to investigate the impact of cognitive and/or motor dual-task

walking conditions on gait dynamics measures in adults with cognitive impairment.

2.4 METHODS

-20-

Protocol and registration

The protocol was prospectively registered with PROSPERO under registration number

CRD42018105787 at: http://www.crd.york.ac.uk/PROSPERO/ on 17/09/2018 date.

Eligibility criteria

Studies were eligible if they met the following criteria: (a) participants subjected to both single-

and dual-task walking conditions within a cross-sectional study design, or at baseline for a

longitudinal observational or experimental study design; (b) full-length article was published in

a peer reviewed journal or an unpublished thesis accessible by reasonable means; (c) human

participants with at least one type of diagnosed objective cognitive impairment, broadly

including mild cognitive impairment (MCI) and dementia, including Alzheimer’s disease (AD),

vascular dementia, frontotemporal dementia, or other variations with the exception of

Parkinson’s Disease or Lewy Body dementia; (d) a cognitive or motor dual-task walking

condition defined as the “simultaneous processing of two (and sometimes more) sources of

information” [22]; (e) a single-task walking condition to allow for isolation of the effects of the

dual-task condition; and (f) one or more objective measures of gait dynamics defined as spatial,

temporal or nonlinear intra-individual stride-to-stride or step-to-step fluctuations in walking

parameters [5].

Studies were excluded if: (a) cognitively impaired participants were grouped with and could not

be separated from participants with: a documented disease with motor effects that had the

potential to impact gait, including but not limited to Parkinson’s disease, Lewy body dementia,

multiple sclerosis (MS), cerebral palsy, spinal cord injury, traumatic brain injury or any other

-21-

disease that has known motor effects on gait; peripheral neuropathy from any cause, including

but not limited to chemotherapy, Charcot-Marie-Tooth (CMT) disease, diabetes mellitus, or

alcoholism; an intellectual disability at birth; cognitive decline due to a delirium or any cause,

psychiatric disorder, medication or other substance use; (b) participants had self-reported

memory concern or a subjective cognitive impairment without a formal diagnosis of cognitive

impairment; (c) participants had undergone one or more above knee amputation (AKA) or below

knee amputation (BKA), or were born without one or part of, their lower limb(s) greater than

that of a single toe; (d) the dual-task gait dynamics outcome of participants with cognitive

impairment was unable to be isolated from participants without cognitive impairment; (e) gait

dynamics outcomes were unable to be extracted from the article. The exclusion of studies of

neurological diseases or musculoskeletal conditions noted above was applied in order to focus

on the cognition or age-related gait dynamics rather than deficits related to these specific

pathologies, which may require different future preventive or therapeutic strategies.

Search strategy

The following electronic databases were selected from the earliest possible date to February

2019: Ageline, CINAHL, EBM review database CCRCTs, EMBASE, Medline, PEDro,

PsychINFO, Scopus and Web of Science. Further, email alerts were set up on all databases and

reviewed weekly until 15th February 2019. To maximize the search sensitivity, the search

strategy included a combination of ‘intervention’ and ‘outcome’ terms, however, it did not

include ‘population’ or ‘comparison’ terms. The intervention terms included ‘dual task*’ OR

‘dual-task*’ OR ‘multi task*’ OR ‘multi-task’ OR ‘secondary task*’ OR ‘attention task*’ OR

‘cognitive task*’ OR ‘motor task*’ OR ‘two task*’ OR ‘2 task*’. The outcome terms included

-22-

walk* OR gait OR locomot* OR ambulat* OR stride* OR step* OR ‘double limb’ OR ‘double-

limb’ OR ‘double support’ OR ‘swing time’ OR ‘stride-to-stride’ OR ‘stride to stride’ OR ‘foot

clearance’ OR mobility OR stability OR instability OR ‘gait variability’ OR ‘centre of pressure’

OR ‘center of pressure’ OR COP OR ‘centre of mass’ OR ‘center of mass’ OR COM OR ataxia

OR McRoberts OR ‘Gait Up’ OR APDM OR GAITRite OR Axivity OR AX3 OR Opal* OR

Pedar OR Zeno OR Gyroscope* OR Lyap* OR fractal. The coding of the search strategy for

each database was customized to search multipurpose (.mp) and the intervention and outcome

searches were then combined with ‘AND’ to produce the final results pool. No language or date

restrictions were applied to the search strategy. A full electronic search strategy example is

presented in Table 2.1. Further eligible trials were hand-searched from the reference lists of all

eligible studies and relevant reviews. Where necessary, authors were contacted for full text

articles or complete gait dynamics datasets in order to identify potential additional studies or

clarify data extracted.

Study selection

One reviewer (TH) performed the literature search and study selection process, which included

the removal of duplicate articles and articles with irrelevant titles or abstracts. The full texts of

the remaining articles were read (TH and TV), with all articles that did not meet the inclusion

criteria removed, and reasoning documented. Any disagreements were resolved by consensus

with other authors (YM and MFS). Remaining eligible articles were included in the systematic

review. In the case that 2 papers were published with the same data set, the paper that was first

published was included [23, 24].

-23-

Data collection process

TH extracted data from each eligible article into pilot-tested collection tables. The data were

verified by direct comparison to the original article (TV). Any discrepancies or disagreements

in data were reviewed and resolved by consensus prior to tabulation (MFS). The data extraction

tables were subsequently refined for increased readability for final manuscript publication. In

the case that multiple walking trials for the same walking condition were conducted, the data

that were collected first were included for analysis to be consistent with studies that included

only 1 trial [25]. In the case of a longitudinal trial with repeated testing, only the baseline data

were used for analysis [26, 27]. In the case that there was more than 1 “healthy” comparison

group, the group that was most closely matched in clinical characteristics (e.g., age, clinical

status) to the cognitive impairment group was included for analysis [28, 29].

Risk of bias assessment

Risk of bias and quality of the included articles were independently appraised by 2 reviewers

(TH and TV) using a modification of the 27-point Downs and Black checklist [30]. Appraisal

was based on reporting, internal validity (bias and confounding) and external validity. Papers in

this review were likely to investigate dual-task walking as an acute exposure. Due to the nature

of this review, items related to follow up (items 9, 17, 26) were not relevant and not included for

scoring. Additionally, item numbers 5, relating to confounder distribution, and 27, relating to

statistical power, were modified to be consistent with the scoring of other items (i.e., alteration

from a 0 to 2 scale, and a 0 to 5 scale, to a 'no, 0; unable to determine, 0; and yes, 1'). Criterion

23 was altered for increased specificity to read: ‘Was the order of the walking tasks (single and

dual) randomized for study subjects?’. To determine if studies were sufficiently powered, a

-24-

clinically significant dual-task cost was defined with reference to the paper by Springer et al.

[31] which reported the difference in dual-task cost between fallers and non-fallers. Therefore,

the maximum modified Downs and Black score possible was 24, with a higher score indicating

better quality.

Summary measures

Primary outcome measures included the objective measurement of spatial, temporal or nonlinear

measures of gait dynamics of stride width, stride time, foot clearance, swing time, stance time,

inconsistency of variance and other aspects of gait dynamics. Gait dynamics could be measured

by any device that recorded intra-individual stride-to-stride fluctuations in walking parameters

[5]. Outcomes may have been expressed as linear measures of gait dynamics such as a co-

efficient of variation (CV) or standard deviation (SD) of stride-to-stride or step-to-step

fluctuations, or nonlinear measures such as approximate entropy (ApEn) or detrended fluctuation

analysis (DFA), which quantify gait irregularity and unpredictability over time-series data [32].

Adverse events related to the dual-task paradigm were extracted if reported, such as falls.

Synthesis of results and Analyses

Studies were split into 3 groups; (a) studies without a control comparison group; (b) studies that

allowed for a comparison between adults with MCI or dementia with adults with a cognitively

healthy control comparison group; and (c) studies that allowed for a direct comparison between

adults with dementia and adults with MCI.

-25-

The primary outcome of interest for this review was gait dynamics as defined above. Gait speed,

number of new falls in the past 12 months (or any shorter time frame) and number of injurious

falls in the past 12 months (or any shorter time frame), visual impairment, hearing impairment,

orthostasis, physical fitness such as aerobic capacity, strength, balance or functional capacity,

depressive symptoms, self-efficacy, nutritional status and quality of life were extracted as

secondary outcomes if reported.

The main data extracted were any quantifiable effects noted on the primary outcome of gait

dynamics, including means and standard deviations (SD) or other summary statistics as

appropriate to the data. The data extracted were at the aggregate level of each study. In addition

to extraction of means and standard deviations (SD) for gait dynamics outcomes, mean

differences (MD) and 95% confidence intervals (95% CI) were calculated and effects sizes (ES)

were calculated as standardized mean difference (SMD) and 95% CI. For studies without a

comparison group, the MD was calculated by subtracting the mean in the single-task condition

from the mean the dual-task condition. The SMD was then calculated by dividing this MD by

the single-task SD using Review Manager 5.3 software (RevMan, Version 5.3; The Nordic

Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark). For studies that had a

comparison group, the MD was calculated by subtracting the mean change from single-to-dual-

task in the control condition from the mean change in single-to-dual-task in the cognitive

impairment condition. The SMD for studies with comparison groups was then calculated by

dividing by the pooled single-task SD. All SMD’s were adjusted for small sample bias (Hedges’

g) [33]. For Tables 2.5, 2.6 and 2.7 the signs of the ESs were reversed in some cases so that a

positive value represented a worsening of gait dynamic measure. Effect sizes were categorized

-26-

according to Cohen’s interpretation of ‘trivial’ (<0.20), ‘small’ (>2.0 to <0.50), ‘moderate’

(>0.50 to <0.80) and ‘large’ (>0.80) [34].

Random effects meta-analyses were attempted for all measures of gait dynamics, with ESs

pooled when I2 was less than 75% using Review Manager (RevMan, Version 5.3; The Nordic

Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark). The I2 range was

chosen to reflect the Cochrane Handbook for Systematic Reviews of Interventions [35]

interpretation of heterogeneity, where an I2 measures of 75-100% reflects ‘considerable

heterogeneity’. For the purposes of the meta-analyses only, when multiple outcomes of gait

dynamics where reported for a group of participants, or a single group had multiple comparisons

to other groups, the total sample size for that group divided by the number of

outcomes/comparisons for that group to account for dependency was used, as recommended by

Cochrane [35]. First, meta-analyses were attempted within the 3 groups (i.e., single- vs. dual-

tasking in MCI or dementia with no cognitively intact control comparison group; cognitively

impaired adults compared to cognitively health adults; and adults with MCI compared to adults

with dementia). Only the direct comparison between adults with MCI and adults with dementia

showing sufficient homogeneity to be pooled (I2= 0%). Attempts to reduce heterogeneity

sufficiently (I2<75%) in the other two groups were ultimately unsuccessful. These steps included

stratifying by cognitive status (MCI or dementia), followed by selecting the most commonly

used measurement of gait dynamics (stride time CV), and finally by stratifying by the cognitive

task. Studies were then removed, and the change in I2 and Q noted, with the study resulting in

the largest reduction removed. As the resultant I2 was not <75%, then this same process was

repeated and a second study removed. However, the I2 still showed considerable heterogeneity

-27-

(I2>75%), and so a meta-analysis was not performed, and a narrative review of the results from

these groups is provided. Results were separated by linear and nonlinear methods of assessing

gait dynamics. Due to the high number of studies reporting coefficient of variation in stride time

as an outcome, stride time CV was reported separately from other linear measures of gait

dynamics, and the MD was reported together with ES for ease of interpretation of clinical

meaningfulness of results. Otherwise, ES was used when reporting studies that used different

outcomes (e.g., swing time CV and stride regularity). Summary results in the narrative synthesis

are presented as MD (range) and ES (range).

2.5 RESULTS

Study selection

The initial keyword search returned 16, 519 results. Following the removal of duplicates, and

title and abstract exclusions, 49 full-text articles were evaluated (Figure 2.1). A further 25 articles

were excluded on the basis of eligibility. In 4 studies where complete data were unable to be

extracted the corresponding authors were contacted. No response was received from 3 studies

[36-38] (n = 3), with a response from 1 study [39], which was then included, for a total of 25

eligible studies.

Study design

In the 25 studies included there were 3 distinct study designs: cross-sectional, longitudinal

observational and longitudinal experimental. This included 10 studies with more than 1 cognitive

impairment comparison group, 9 studies with more than 1 dual-task procedure and 9 studies with

-28-

more than 1 gait dynamics outcome measured. The cohort characteristics of the included studies

are displayed in Table 2.2.

Quality

Modified Downs and Black scores are shown in Table 2.3. The average study quality was

moderate, 13.6 (range: 11 – 17/24). The most common limitations were lack of subject and

assessor blinding, adverse event reporting, reporting of participant representativeness within

population and sample, recruited participant source population and representativeness,

recruitment time frame, and concealment of intervention/procedure. It is acknowledged that

blinding of participants and assessors, and the concealment of intervention/procedure to

participants were not possible due to the study designs, thus potentially limiting the maximal

score to 21 rather than 24. Additionally, as most study designs were acute exposure, the

opportunity for an adverse event to occur was reduced, which potentially limited the reporting

of such events. Lack of reporting of participant representativeness within the population and

sample, however, are threats to external study validity and were deficient in most studies.

Cohort characteristics

Cohort characteristics are presented in Table 2.2. Across all studies 1118 participants (56.5%

female) were included, 797 cognitively impaired (55.2% female) and 321 cognitively healthy

(59.8% female). Twenty-three studies (92.0%) included both male and female participants and

2 studies (8.0%) [40, 41] did not specify sex. The mean reported age for all included study

participants was 76±5 years (range: 59-94 years), with the mean reported age for the cognitively

impaired 77±6 years (range: 61-92 years) and for the cognitively heathy 74±4 years (range: 59-

-29-

94 years). Habitual gait speed was reported in 19 studies [8, 11, 23, 25-29, 39-49] for cognitively

impaired participants (0.94±0.25 m/s) and 9 studies [11, 23, 28, 29, 40, 42, 45-47] for cognitively

healthy participants (1.11±0.16 m/s).

Studies included participants with varying cognitive impairment diagnoses: 14 studies (52.0%)

[25, 26, 28, 41-43, 45, 46, 48-53] included dementia only, 8 studies (32.0%) [8, 10, 27, 29, 39,

40, 44, 47] included MCI only and 3 studies (12.0%) [11, 23, 54] included both dementia and

MCI. Fourteen studies (56.0%) [10, 11, 23, 28, 29, 40, 42, 45-47, 50, 52-54] included a

cognitively healthy control group as a comparison group. The Winblad [55] criteria and the

Petersen [12] criteria were the most common MCI diagnostic criteria used, cited by 4 [10, 11,

27, 29] and 3 studies [8, 44, 54], respectively. The National Institute of Neurological and

Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders

Association (NINCDS-ADRDA) [56] and the Diagnostic and Statistical Manual of Mental

Disorders (DSM-IV) (4th edition) were the most common dementia diagnostic criteria, cited by

6 [23, 26, 41-43, 49] and 3 studies [26, 42, 50], respectively. Numerous assessments were used

to determine the severity of MCI and dementia, with the Mini-mental State Exam (MMSE) the

most common. Twenty-two studies [8, 10, 11, 23, 25, 26, 28, 39-46, 48-54] reported MMSE

scores for cognitively impaired participants (21.6±4.1), specifically MCI (26.4±1.1) and

dementia participants (19.9±0.9), and 12 studies [10, 11, 23, 28, 40, 42, 45, 46, 50, 52-54]

reported MMSE scores for cognitively healthy participants (28.8±1.1).

Measurement of gait dynamics

Characteristics of dual-task procedures are presented in Table 2.4. All studies included flat

ground walking and participants were instructed to walk at their usual pace. The mean measured

-30-

walking distance was 21.66m (range: 3.66-160m). Twelve studies [8, 10, 11, 25-27, 43, 48-52]

used the GAITRite system (CIR Systems Inc., Clifton, NJ, USA) to measure gait dynamics, 4

studies [41, 42, 47, 52] used footswitch sensors, 3 studies [23, 39, 44] used Locometrix (Centaure

Metrix, Evry, Essonne, France), 2 studies [28, 46] used DynaPort MiniMod (McRoberts BV,

The Hague, The Netherlands), 2 studies [40, 53] used motion capture cameras and 3 studies [29,

45, 54] used movement tracking devices. The mean number of trials per walking condition was

1.7 (range: 1-6 trials), with 2 studies [23, 43] collecting data from 1 pre-nominated trial when

multiple trials were completed per condition, (e.g., data were collected for the second trial out of

the 3 conducted trials per condition), and 7 studies [8, 10, 11, 29, 41, 45, 50] not reporting the

number of trials completed. The order of single-task and dual-task walking was inconsistent

between studies; 9 studies [8, 10, 11, 27, 42, 47, 50, 51, 53] completed the tasks in a random

order, 7 studies [25, 39-41, 43, 44, 49] completed the tasks in a set order, (i.e., single-task then

dual-task), and 9 studies [23, 26, 28, 29, 45, 46, 48, 52, 54] did not report the task order. No

studies with a non-randomized task order corrected their analytical models for order effects.

Dual-task procedure characteristics

Characteristics of dual-task procedures are presented in Table 2.4. Dual-tasks were categorized

into two types: cognitive and motor. Twenty-four studies implemented cognitive protocols [8,

10, 11, 23, 25-29, 39-48, 50-54] and 2 studies implemented motor protocols [48, 49]. The

cognitive protocols were further separated into 4 sub-types [16]: mental tracking, verbal fluency,

working memory and verbal memory. Mental tracking, the task of holding information mentally

while manipulating the same information, to measure sustained attention [57], was carried out

in 21 studies [8, 10, 11, 23, 26, 27, 29, 39, 40, 42, 44, 45, 47, 48, 50-54]. All 21 studies reported

-31-

backwards counting, however, the protocol varied between studies with the starting number

ranging from 30 to 378 and the counting increment ranging from 1 to 7. The most common

protocols were backward counting from 50 and 100 by 1s, carried out by 7 [23, 26, 39, 42, 44,

50, 52] and 5 studies [8, 10, 11, 29, 45], respectively. Verbal fluency, the task of spontaneously

producing words within specific constraints to measure executive function [57], was carried out

in 12 studies [8, 10, 11, 25, 27-29, 43, 46, 47, 51, 53]. Two verbal fluency protocols were

reported: animal naming and categorical letter naming, which 5 studies [8, 10, 11, 27, 29] and 3

studies [28, 46, 47] carried out. Working memory, the task of holding information mentally for

later processing [58] was carried out in 5 studies [25, 41, 43, 51, 53] and included forwards

counting [25, 43, 51] and forward digit span [41, 53] protocols. Forwards counting and forward

digit span were included as working memory tasks, rather than mental tracking tasks, as forward

recall loads onto a separable short-term memory factor unlike backward counting and digit span,

which require a more attention-demanding transformation of the digit sequence [59]. Verbal

memory, the task of recalling specific past events or information using speech [60], was carried

out in 1 study [47] in the form of short story recall.

Gait dynamics outcomes

Gait dynamics outcomes are presented in Table 2.5, 2.6 and 2.7. Twenty-six spatiotemporal gait

dynamics outcome measures were examined across all studies. Stride time CV, step length CV

and stride regularity were the most commonly measured, with 18 studies [8, 10, 11, 25-28, 40-

42, 45-47, 49-53], 4 studies [25, 48, 49, 53] and 3 studies [29, 39, 44] reporting each outcome,

respectively.

-32-

Single- vs. dual-tasking in MCI or dementia with no cognitively intact control comparison group

Among the 25 studies, 11 studies had at least 1 cohort with MCI and 17 studies had at least 1

cohort with dementia, resulting in 84 ESs comparing single- to dual-tasking (Table 2.8). A meta-

analysis was attempted but was not appropriate due to considerable heterogeneity (I2=85%,

Tau2=1.32, Q=553.30, p<0.0001). Attempts to reduce heterogeneity included stratifying by

cognitive status (MCI or dementia), and then further stratifying by gait dynamics outcome and

cognitive task, but were ultimately unsuccessful. Therefore, a narrative synthesis of ESs is

presented.

Stride time CV

In cohorts with MCI, stride time CV was the most commonly used outcome of gait dynamics,

resulting in 17 ESs from 6 studies (Table 2.5a). Ten out of the 17 ESs involved the addition of a

mental tracking cognitive dual-task (5 counting backwards by 1; 1 counting backward by 2; and

4 counting backwards by 7), 6 used a categorical verbal fluency dual-task (5 used animal naming

and 1 used letter specific word naming), while the remaining ES was verbal memory in the form

of a short story recall. Overall, the data showed that the addition of a cognitive dual-task resulted

in a significant increase in stride time CV, with 15 of 17 ESs reported being significant (MD

ranging from 1.42% to 8.77%). The 2 non-significant results were from the same study [8], which

used backwards counting by 1, with the subgroup with non-amnestic MCI unexpectedly

achieving a smaller dual-task cost with the addition of a cognitive distractor. Overall, the median

MD was 2.43% (range: 0.50% to 8.77%), while ES was 1.87 (range: 0.35 to 7.70).

-33-

Similarly, in cohorts with dementia, stride time CV was the most commonly used measure of

gait dynamics, resulting in 20 ESs from 13 studies (Table 2.5b). Eleven out of the 20 ESs

involved the addition of a mental tracking cognitive dual-task (8 counting backwards by 1; 1

counting backward by 7; 1 counting backwards by an unspecified number; and 1 backward 3-

digit span), 5 used a working memory cognitive task (2 counting forwards by 1; 1 counting

forwards by an unspecified number and 2 forward 3-digit span), 3 used a categorical verbal

fluency dual-task (1 used animal naming and 2 used letter specific word naming), while the

remaining ES was a motor dual-task of tray carrying. Overall, the data showed that the addition

of a cognitive dual-task resulted in heterogeneous effects on stride time CV, with 12 significant

(MD ranging from 2.19 to 29.00) and 7 non-significant ESs (MD ranging from -0.53% to 3.00%).

Three non-significant studies [28, 41, 52] still reported potentially clinically meaningful MD

changes in dual-task cost (MDs of 2.6%, 2.6% and 3.0% for stride time CV), suggesting the

possibility of type II error, with respective ESs of 0.74 (95% CI -0.03, 1.51), 0.97 (95% CI -0.16,

2.10) and 0.53 (95% CI -0.50, 1.57). Overall, the median MD was 2.97% (range: -0.53% to

29.00), while ES was 1.26 (range: -0.18 to 42.18).

Other linear measures of gait dynamics

In cohorts with MCI, 11 ESs from 5 studies were calculated using other linear measures of gait

dynamics (Table 2.5a). These included step regularity (n=3), stride regularity (n=5), step time

CV (n=2) and step time variance (n=1). Results were heterogeneous with 6 significant ES

(ranging from 1.03 to 26.65). Overall, the median ES for all other linear measures of gait

dynamics was 1.07 (range: 0.24 to 2.74).

-34-

In cohorts with dementia, 14 other methods of linear gait dynamics were used across 9 studies,

resulting in 34 ESs (Table 2.5b). Twenty-nine of these ESs were generated from relatively simple

cognitive dual-tasks that involved either forward counting by 1 (n=16), backward counting by 1

(n=11), forward 3-digit span (n=1) or backward 3-digit span (n=1). Results were mostly

negative, suggesting little difference between single- and dual-task gait dynamics, with 22 of 34

ESs non-significant, and the median ES for all other linear measures of gait dynamics 0.51

(range: -0.40 to 45.71).

Nonlinear measurements of gait dynamics

Only 2 studies used nonlinear measures of gait dynamics, resulting in 3 ESs. Gait dynamic

outcomes for cohorts with MCI and dementia are presented in Tables 2.5a and 2.5b, respectively.

One study was in adults with MCI (2 ESs), while the remaining study was in adults with

dementia. In adults with MCI, gait dynamics were assessed using ApEn using backwards

counting by 1 and animal naming as the cognitive distractors, with both resulting in very large,

significant worsening of gait [ES = 14.93 (95% CI 7.05, 22.81) for backwards count by 1, and

ES = 26.97 (95% CI 12.85, 41.09) for animal naming]. By contrast, in adults with dementia, gait

dynamics were measured using DFA, with a word naming verbal fluency task, showing no

change in gait dynamics during the dual-task [ES = 0 (95% CI -0.19, 0.19)].

Cognitive impairment vs. control comparison group

Fourteen studies included a healthy control group; 7 MCI studies [10, 11, 23, 29, 40, 47, 54] and

10 dementia studies [11, 23, 28, 42, 45, 46, 50, 52-54] resulting in 40 ESs. A meta-analysis was

attempted but was not appropriate due to considerable heterogeneity (I2=100%, Tau2=3.35,

-35-

Q=20054.93, p<0.00001). Attempts to reduce heterogeneity were unsuccessful, which included

stratifying by cognitive status (MCI or Dementia), as well as stratifying by gait dynamics

outcome and cognitive task, results summarized in Table 2.8). Therefore, a narrative synthesis

of ESs is presented.

Stride time CV

In cohorts with MCI, stride time CV was the most commonly used outcome of gait dynamics,

resulting in 9 effect sizes from 4 studies. Five out of the 9 ESs involved the addition of a mental

tracking cognitive dual-task (3 counting backwards by 1; 2 counting backwards by 7), 3 used a

categorical verbal fluency dual-task (2 used animal naming; 1 used letter specific word naming),

while the remaining ES was verbal memory in the form of a short story recall. Overall, the data

showed that the dual-task cost of a cognitive dual-task for adults with MCI compared to

cognitively healthy controls resulted in a significant increase in stride time CV, with 8 of 9 ESs

reported significant (MD ranging from 1.07% to 7.05%). Gait dynamic outcomes for cohorts

with MCI compared to cognitively healthy controls are presented in Table 2.6a. Overall, the

median MD was 2.75% (range: 0.26% to 7.05%) and ES 2.43 (range: 0.16 to 6.53).

Similarly, in cohorts with dementia compared to cognitively healthy controls, stride time CV

was the most commonly used measure of gait dynamics, resulting in 13 ESs from 8 studies. Nine

out of the 13 ESs involved the addition of a mental tracking cognitive dual-task (7 counting

backwards by 1, 1 counting backward by 7, and 1 backward 3-digit span), 3 used a categorical

verbal fluency dual-task (1 used animal naming and 2 used letter-specific word naming), while

the remaining ES was a working memory cognitive task (forward 3-digit span). Overall, the data

-36-

showed that the addition of a cognitive dual-task for adults with dementia compared to

cognitively healthy controls resulted in heterogeneous effects on stride time CV, with 9 ESs

showing worsening of gait (MD ranging from 1.79% to 16.64%), 1 ES showing an improvement

in gait (MD -2.58%) and 3 non-significant ESs (MD ranging from (-0.40% to 2.25%). Gait

dynamic outcomes for cohorts with dementia compared to cognitively healthy controls are

presented in Table 2.6b. Overall, the median MD was 2.23% (range: -2.58% to 16.64%) and ES

1.97 (range: -1.09, 22.03).

Other linear measures of gait dynamics

In cohorts with MCI, 8 effect sizes from 3 studies were calculated for other linear measures of

gait dynamics, including step regularity (n=3), stride regularity (n=2), step time CV (n=2) and

step time variance (n=1). Six of the 8 ESs were significant. Non-significant results used the

simple task of backwards counting by 1 (n=2) only. Gait dynamic outcomes for cohorts with

MCI compared to cognitively healthy controls are presented in Table 2.6a. Overall, the median

ES for all other linear measures of gait dynamics was 0.63 (range: 0.23, 1.36).

In cohorts with dementia, 8 ESs from 4 studies were calculated from linear measures of gait

dynamics. The 8 ESs were generated from cognitive dual-tasks of backwards counting by 1

(n=6), backwards 3-digit span (n=1) or forward 3-digit span (n=1). Results were heterogeneous

with 6 of 8 ESs significant but varying greatly in magnitude (ranging from 0.66 to 23.20). Gait

dynamic outcomes for cohorts with dementia compared to cognitively healthy controls are

presented in Table 2.6b. Overall, the median ES for all other linear measures of gait dynamics

was 1.20 (range: -0.10 to 23.2).

-37-

Nonlinear measurements of gait dynamics

Only 2 studies used nonlinear measures of gait dynamics, resulting in 3 ESs. Gait dynamic

outcomes for cohorts with MCI compared to cognitively healthy controls and dementia

compared to healthy controls are presented in Table 2.6a and Table 2.6b, respectively. Two ESs

came from 1 study [29] in adults with MCI, while the remaining ES was from a study of adults

with dementia [46]. In adults with MCI, gait dynamics were assessed via ApEn using animal

naming and backwards counting by 1 as the cognitive distractors. Neither distractor resulted in

a significant difference in dual-task cost in the MCI cohort [ES = 0.45 (95% CI -0.29 to 1.20)]

and [ES = -0.62 (95% CI -1.37 to 0.14)], respectively. By contrast, in adults with dementia

compared to cognitively healthy controls, gait dynamics were measured using DFA, with a word

naming verbal fluency task, showing a significant worsening of DFA during dual-tasking [ES =

0.81 (95%CI: 0.01, 1.62)].

MCI vs. dementia groups

Three studies included both MCI and dementia groups [11, 23, 54] and measured gait outcomes

using stride time CV [11], step regularity [23] and step time variance [54]. Gait dynamic

outcomes for cohorts with MCI vs. dementia are presented in Table 2.7. A meta-analysis was

performed (Figure 2.2), showing that adults with dementia have a moderate, significant

worsening of gait dynamics under dual-task conditions compared to adults with MCI [ES = 0.60

(95%CI: 0.22, 0.99), I2=0%, Tau2=0.00, Q=3.74, p=0.002).

Adverse events

-38-

Only 1 study [27] reported adverse events, however this study was a placebo-controlled drug

study where none of the reported events were attributed to baseline single- or dual-task

procedures. No study reported whether single- or dual-task walking was associated with any

adverse event such as falling.

Other cohort characteristics

Limited information was available on cohort characteristics other than gait or cognitive

performance, with fall history and measures of physical fitness being the most often reported

characteristics. Six studies [8, 10, 25, 42, 48, 49] reported falls history in the past 6 to 12 months.

A sensitivity analysis adjusting for history of falls and age was conducted in 1 study [10], the

direction and magnitude of the dual-task costs were maintained (a difference <10% from the

unadjusted values). A comparison between multiple fallers and non-multiple fallers was

conducted in 1 study [48], which showed that there was no significant interaction between dual-

task cost and faller status. One study [42] adjusted for previous falls and other baseline

characteristics using a multivariate linear regression for single-task and dual-task conditions and

stride time CV, which showed that previous falls were not significantly related to stride time CV

under these conditions. However, no statistical analysis with respect to dual-task cost of gait

dynamics reported. A statistical analysis by faller status with respect to dual-task cost of gait

dynamics was not reported in 3 other studies [8, 43, 49]. History of falls was listed as an

exclusion criterion in 4 studies [11, 23, 40, 44]. Additionally, 2 studies [8, 10] reported fear of

falling, with no statistically significant between-group comparisons in either study (p=0.77 [10]

and p=0.84 [8]) and no statistical analysis with respect to the dual-task cost of gait dynamics was

reported.

-39-

Physical activity or functional mobility assessments were reported in 6 studies [8, 10, 11, 23, 29,

39], which included self-reported physical activity level, one-legged balance, and Timed Up and

Go (TUG) test. Three studies [8, 10, 11] reported self-reported physical activity level. One study

[8] adjusted for physical activity level using a multivariable linear regression, and, as previously

reported, amnestic MCI participants had statistically significantly higher stride time CV (p=0.01)

compared to non-amnestic participants under single- and dual-task walking after adjustment. No

other studies reported relating these characteristics to gait performance.

Nutritional status was measured in 2 studies [23, 44] using the Mini-Nutritional Assessment

(MNA), however, no statistical analysis with respect to the dual-task cost of gait dynamics was

reported. Five studies [23, 26, 27, 39, 44] reported that depression was an exclusion criterion,

measured by the Geriatric Depression Scale (GDS) [61] or the Hospital Anxiety Depression

Scale (HADS) [62], but these scores were not included in models of gait dynamics. Injurious

falls in the past 12 months (or any shorter time frame), orthostasis, self-efficacy and quality of

life were not reported in any study.

2.6 DISCUSSION

To our knowledge, this is the first systematic review to solely investigate the effects of dual-task

walking on gait dynamics in older adults with cognitive impairment. The purpose of the review

was to evaluate the effects of dual-task walking on changes in gait dynamics and dynamics with

respect to pathology diagnosis, dual-task paradigm and outcome measure in older adults with

cognitive impairment. In total, 27 different spatial, temporal and nonlinear measures of gait

-40-

dynamics and 20 different dual-task procedures were identified in the literature. The overall

findings of this review are: 1) gait dynamics are worse under dual-task conditions than single-

task conditions in older adults with MCI and dementia; 2) this dual-task cost is greater in

cognitively impaired older adults than in healthy older adults; 3) when MCI and dementia are

directly compared, dual-task cost is greater in older adults with dementia; and 4) characteristics

associated with the dual-task cost of gait dynamics in cognitively impaired older adults have

been minimally investigated.

This review indicates that the addition of a dual-task while walking alters gait dynamics in older

adults with cognitive impairment. This is in agreement with previous studies in healthy older

adults [63], Parkinson’s disease [64], multiple sclerosis [65] and MCI [19]. As expected, dual-

task gait dynamics, measured by stride time CV, increased during mental tracking tasks and

verbal fluency tasks in both older adults with MCI and dementia. This directly reflects outcomes

in healthy older adults, where significant increases in dual-task compared to single-task stride

time CV have been reported for both backwards counting and verbal fluency tasks [66]. As

observed, a more complex task (i.e., dual-tasking compared to single-tasking), produced more

cognitive interference, which resulted in larger variability and a greater cost to gait dynamics

[10, 49]. This increase in dual-task cost is concordant with the literature, which shows that dual-

tasking predicts falls risk better than single-task gait measurements in healthy older adults,

however it has not been as clear in cognitively impaired older adults [9, 67]. If ways are identified

to minimize the cost of dual-tasking on gait dynamics, the frequency and severity of future falls

could potentially be reduced. However, very little data are available to indicate what the

potentially modifiable contributants to dual-task costs are in cognitively impaired cohorts. We

-41-

searched for potential candidates such as low physical fitness, depression, low physical activity

levels, undernutrition, or fear of falling, but unfortunately there was minimal reporting of any

such characteristics, nor of their effect on dual-task cost. Future studies should include such

characterization, and assess relationships of relevant factors to gait outcomes under single- and

dual-task conditions to advance this field.

The dual-task cost of gait dynamics was larger in cognitively impaired older adults than in

cognitively healthy older adults. Specifically, a greater decrement in dual-task gait dynamics was

observed during mental tracking and verbal fluency tasks in both older adults with MCI and

dementia compared to healthy controls. Gait dynamics are maintained and stable with age in

healthy older adults [68], despite dual-tasking, which differs from the significant worsening

under dual-task conditions reported in cognitively impaired older adults [11], and confirmed by

this review. Physiological and pathological aging impact gait ability and cognitive function, and

the association between these two factors suggests that a complex age-related relationship exists

[66]. Broadly, poor gait performance has been identified as predictive of dementia, with a

stronger association in non-AD dementias than in AD [69]. Specifically, increases in stride time

CV under dual-task conditions have been shown to predict cognitive decline [66] suggesting that

loss of gait stability may be an early sign of brain pathology. Cognitive impairment and executive

function impairment are both associated with an increased risk of falls, while global measures of

cognition are associated with serious fall-related injury [14]. Thus, dual-task gait testing has been

recommended as part of the assessment battery to determine risk of falls in all cognitively

impaired older adults [11]. In addition, it may identify individuals with subtle cognitive changes

-42-

who may benefit from detailed evaluation of cognition and assessment for potentially treatable

aetiologies.

The dual-task decrement was larger for individuals with greater cognitive impairment, i.e.,

dementia compared to MCI. A meta-analysis was performed on the 3 studies [11, 23, 54] that

included both MCI and dementia groups. Individuals with dementia had a significantly and

moderately increased dual-task cost compared to individuals with MCI. Previous studies,

including those analyzed, have reported conflicting results as to whether there is a significant

difference between dual-task gait performance in older adults with MCI and AD [11, 23, 24, 54,

70]. Evidence suggests that older adults with AD are slower on basic mobility tests [70], have

slower gait speed [71], perform worse on cognitive tasks [72], while it has been previously shown

that adults with AD and dementia demonstrate increased gait dynamics during single-task

walking than older adults with MCI [71], and thus a ceiling effect may explain some of the

heterogeneity between studies. Additionally, older adults who are more impaired may potentially

be limited in their ability to adopt protective strategies during dual-tasking (e.g., increasing step

length to compensate for gait abnormalities or dysfunction) [73], however, few comparative

studies exist that investigate potential changes across the cognitive spectrum [11], and the gait

measures within the 3 included studies [11, 23, 54] were inconsistent. When clinically assessing

dual-task gait deficits it is important to consider individual characteristics such as the severity of

motor and cognitive impairments, concurrent tasks complexity and the environmental challenge

on the falls risk [73]. To understand more about the impact of the dual-task on gait dynamics

across the cognitive continuum, additional studies are required to examine gait using comparable

outcome measures and dual-tasks.

-43-

The included studies were too inconsistent in their methodology and reporting to determine if a

dual-task type or a specific gait dynamic measure was better at producing or detecting dual-task

cost. Additionally, only two studies measured gait dynamics using nonlinear outcomes (i.e., DFA

[46] and ApEn [29]). The outcomes for DFA and ApEn are not directly comparable, with the

two outcomes being measured in different cohorts using different dual-task conditions.

Additionally, it is not expected to observe the same outcome for different nonlinear outcomes,

with DFA measuring the degree of randomness in highly non-stationary data and ApEn

measuring the likelihood that a template pattern repeats in a time series [74]. In order to have a

complete understanding of the dynamics of gait, both linear and nonlinear measures are needed.

Further research is required using linear and nonlinear outcomes as the primary measure of

interest, with respect to dual-task walking conditions of varying complexity, in order to enhance

the understanding of falls risk and translation of findings into research and clinical guidelines.

No specific characteristics related to the dual-task gait dynamics in cognitively impaired older

adults were identified. However, the reporting of any cohort characteristics relating to falls were

poor, with few studies documenting history of falls, fear of falling or tracking falls over time. In

these studies, individuals who had fallen more times were likely to have worsened dual-task gait

dynamics than those who had fallen fewer times, however, this did not translate into a

significantly increased dual-task cost [43, 48]. Dual-task gait changes are associated with future

fall risk [9], however, the link between future risk of falls and increased gait dynamics is poorly

studied in cognitive impairment. Previous research has shown a greater dual-task cost of gait

dynamics is associated with progressive cognitive decline and an increase in falls risk, although

-44-

this association was not reported in this review [75]. Few studies reported other characteristics

such as physical activity level static balance, functional mobility, nutritional status and

depression, while no studies reported aerobic capacity, strength, self-efficacy, or quality of life.

If these or other modifiable characteristics were able to be identified, a targeted intervention

could be implemented to potentially reduce the risk of falls in this population [49].

2.7 STRENGTHS

The strengths of this review were that it included a broad, sensitive search strategy across all

years and major databases, resulting in a large number of retrieved citations. Additionally, the

analysis of results was stratified by type of cognitive impairment and dual-task condition to

attempt to create more uniformity in the interpretation of dual-task gait dynamics methodology.

This review does not duplicate previous work and reflects the current literature in this topic area

allowing it to help drive recommendations for knowledge translation and act as a reference point

for future research strategies.

2.8 LIMITATIONS

This review was limited by the use of only one author responsible for the initial study selection

and data extraction. However, full text inclusion and exclusion were independently performed

by two authors, and consensus was obtained by a third author. Additionally, unpublished data

were neither searched for nor included. Furthermore, this review was restricted to one meta-

analysis of three studies due to all other groupings displaying an I2 value greater than 75%. The

I2 value represents the consistency of study results and assesses whether differences in results

between studies are compatible with chance alone. Cochrane categorize an I2>75% as possessing

-45-

considerable heterogeneity, which was used to support the decision of using 75% as a cut off

value. While the small number of studies in the meta-analysis may require caution with regards

to interpretation, the direction of the effect of this analysis was in agreement with all other data

presented within this review. Several attempts were used to reduce the heterogeneity, but were

ultimately unsuccessful. Future studies should attempt to identify factors contributing to

heterogeneity across studies.

2.9 CONCLUSIONS

Gait dynamics worsen under dual-task conditions compared to single-task conditions in

cognitively impaired older adults. The dual-task cost of gait dynamics increases in cognitively

impaired older adults compared to healthy older adults, with worse gait dynamics observed with

a greater degree of cognitive dysfunction. Data are too inconsistent currently to determine which

type of dual-task is best able to expose gait instability, or which measure of gait dynamics best

predicts risk of falls or level of cognitive impairment in older adults. Importantly, only two

studies reported nonlinear gait outcome measures, highlighting an area where more research is

needed to understand the complete impact of dual-tasking on gait dynamics. Additionally, factors

that may impact gait dynamics, including history of falling, fear of falling, physical fitness,

objective physical activity or sedentary behavior, depression, nutritional status, vision, hearing,

overall disease burden and medication use, and quality of life were documented poorly or not at

all. To adequately determine the modifiable and non-modifiable characteristics of dual-task gait

dynamics, more well-designed longitudinal studies and controlled trials with adequately

powered samples are needed. Thus, by identifying characteristics of dual-task gait dynamics,

-46-

clinical interventions could be developed to target these modifiable factors with the aim to reduce

the high falls risk in this population.

-47-

2.10 REFERENCES

1. Callisaya, M.L., et al., Gait, gait variability and the risk of multiple incident falls in older

people: a population-based study. Age and Ageing, 2011. 40(4): p. 481-487.

2. Stewart Williams, J., et al., Prevalence, risk factors and disability associated with fall-

related injury in older adults in low- and middle-incomecountries: results from the WHO

Study on global AGEing and adult health (SAGE). BMC Medicine, 2015. 13(1): p. 147.

3. Lundin-Olsson, L., L. Nyberg, and Y. Gustafson, “Stops walking when talking” as a

predictor of falls in elderly people. The lancet, 1997. 349(9052): p. 617.

4. Bradley, C. and J. Harrison, Hospitalisations due to falls in older people, Australia, 2003-

2004, in Injury research and statistics series no. 32. 2007, AIHW: Canberra.

5. Hausdorff, J.M., Gait variability: methods, modeling and meaning. Journal of

NeuroEngineering and Rehabilitation, 2005. 2(1): p. 19.

6. Hausdorff, J.M., et al., Walking is more like catching than tapping: gait in the elderly as

a complex cognitive task. Experimental Brain Research, 2005. 164(4): p. 541-8.

7. Ayers, E.I., et al., Walking while Talking and Falls in Aging. Gerontology, 2014. 60(2):

p. 108-113.

8. Montero-Odasso, M., et al., The motor signature of mild cognitive impairment: results

from the gait and brain study. Journals of Gerontology Series A-Biological Sciences &

Medical Sciences, 2014. 69(11): p. 1415-1421.

9. Muir-Hunter, S.W. and J.E. Wittwer, Dual-task testing to predict falls in community-

dwelling older adults: a systematic review. Physiotherapy, 2016. 102(1): p. 29-40.

10. Montero-Odasso, M., S.W. Muir, and M. Speechley, Dual-task complexity affects gait in

people with mild cognitive impairment: the interplay between gait variability, dual

tasking, and risk of falls. Archives of Physical Medicine & Rehabilitation, 2012. 93(2):

p. 293-299.

-48-

11. Muir, S.W., et al., Gait assessment in mild cognitive impairment and Alzheimer's disease:

The effect of dual-task challenges across the cognitive spectrum. Gait and Posture., 2012.

12. Petersen, R.C., Mild Cognitive Impairment: Clinical Characterization and Outcome. The

Journal of the American Medical Association, 1999. 281(19): p. 303-308.

13. Petersen, R.C., et al., Current concepts in mild cognitive impairment. Archives of

Neurology, 2001. 58(12): p. 1985-1992.

14. Muir, S.W., K. Gopaul, and M.M. Montero Odasso, The role of cognitive impairment in

fall risk among older adults: a systematic review and meta-analysis. Age and Ageing,

2012. 41(3): p. 299-308.

15. Shaw, F.E., Falls in cognitive impairment and dementia. Clinics in Geriatric Medicine,

2002. 18(2): p. 159-173.

16. Al-Yahya, E., et al., Cognitive motor interference while walking: A systematic review

and meta-analysis. Neuroscience and Biobehavioral Reviews, 2011. 35(3): p. 715-728.

17. Li, K.Z.H., et al., Cognitive Involvement in Balance, Gait and Dual-Tasking in Aging: A

Focused Review From a Neuroscience of Aging Perspective. Frontiers in Neurology,

2018. 9: p. 913.

18. Moon, Y., et al., Gait variability in people with neurological disorders: A systematic

review and meta-analysis. Human Movement Science, 2016. 47: p. 197-208.

19. Bahureksa, L., et al., The Impact of Mild Cognitive Impairment on Gait and Balance: A

Systematic Review and Meta-Analysis of Studies Using Instrumented Assessment.

Gerontology, 2017. 63(1): p. 67-83.

20. Bragatto, V.S.R., et al., Dual-task during gait between elderly with mild cognitive

impairment and Alzheimer: systematic review. Fisioterapia em Movimento, 2017. 30: p.

849-857.

21. Menant, J.C., et al., Single and dual task tests of gait speed are equivalent in the

prediction of falls in older people: a systematic review and meta-analysis. Ageing

Research Reviews, 2014. 16: p. 83-104.

-49-

22. Verhaeghen, P., et al., Aging and Dual-Task Performance: A Meta-Analysis. Psychology

and Aging, 2003. 18(3): p. 443-460.

23. Gillain, S., et al., The value of instrumental gait analysis in elderly healthy, MCI or

Alzheimer's disease subjects and a comparison with other clinical tests used in single

and dual-task conditions. Annals of Physical & Rehabilitation Medicine, 2009. 52(6): p.

453-74.

24. Maquet, D., et al., Gait analysis in elderly adult patients with mild cognitive impairment

and patients with mild Alzheimer's disease: simple versus dual task: a preliminary report.

Clinical Physiology & Functional Imaging, 2010. 30(1): p. 51-56.

25. Camicioli, R., T. Bouchard, and L. Licis, Dual-tasks and walking fast: relationship to

extra-pyramidal signs in advanced Alzheimer disease. Journal of the Neurological

Sciences, 2006. 248(1-2): p. 205-9.

26. Beauchet, O., et al., Gait Changes with Anti-Dementia Drugs: A Prospective, Open-

Label Study Combining Single and Dual Task Assessments in Patients with Alzheimer's

Disease. Drugs & Aging, 2014. 31(5): p. 363-372.

27. Gschwind, Y., et al., Ginkgo biloba special extract LI 1370 improves dual-task walking

in patients with MCI: a randomised, double-blind, placebo-controlled exploratory study.

Aging clinical and experimental research. Issue, 2017. 4: p. 609-619.

28. Ijmker, T. and C.J. Lamoth, Gait and cognition: the relationship between gait stability

and variability with executive function in persons with and without dementia. Gait &

Posture, 2012. 35(1): p. 126-30.

29. Martinez-Ramirez, A., et al., Dual Task Gait Performance in Frail Individuals with and

without Mild Cognitive Impairment. Dementia & Geriatric Cognitive Disorders, 2016.

42(1-2): p. 7-16.

30. Downs, S.H. and N. Black, The feasibility of creating a checklist for the assessment of

the methodological quality both of randomised and non-randomised studies of health

-50-

care interventions. Journal of Epidemiology and Community Health, 1998. 52(6): p. 377-

384.

31. Springer, S., et al., Dual-tasking effects on gait variability: The role of aging, falls, and

executive function. Movement Disorders, 2006. 21(7): p. 950-957.

32. Arif, M., et al., Estimation of the effect of cadence on gait stability in young and elderly

people using approximate entropy technique. Measurement Science Review, 2004. 4(2):

p. 29-40.

33. Hedges, L.V., Distribution theory for Glass's estimator of effect size and related

estimators. journal of Educational Statistics, 1981. 6(2): p. 107-128.

34. Cohen, J., Statistical power analysis for the behavioural sciences. Second edition ed.

1988, Hillside, NJ, USA: Lawrence Erlbaum Associates.

35. Higgins, J. and S.e. Green, Cochrane Handbook for Systematic Reviews of Interventions,

in Version 5.1.0 [updated March 2011]. 2011, The Cochrane Collaboration: The

Cochrane Collaboration.

36. Kikkert, L., et al., Gait characteristics and their discriminative power in geriatric

patients with and without cognitive impairment. Journal of NeuroEngineering and

Rehabilitation, 2017. 14.

37. Montero-Odasso, M., et al., Quantitative gait analysis under dual-task in older people

with mild cognitive impairment: a reliability study. Journal of Neuroengineering &

Rehabilitation, 2009. 6: p. 35.

38. Rucco, R., et al., Spatio-temporal and kinematic gait analysis in patients with

Frontotemporal dementia and Alzheimer's disease through 3D motion capture. Gait &

Posture, 2017. 52: p. 312-317.

39. Auvinet, B., et al., Gait disorders in the elderly and dual task gait analysis: a new

approach for identifying motor phenotypes. Journal of Neuroengineering &

Rehabilitation, 2017. 14(1): p. 7.

-51-

40. Lee, J. and S. Park, Effects of a priority-based dual task on gait velocity and variability

in older adults with mild cognitive impairment. Journal of Exercise Rehabilitation, 2018.

14(6): p. 993-997.

41. Sheridan, P.L., et al., Influence of executive function on locomotor function: divided

attention increases gait variability in Alzheimer's disease. Journal of the American

Geriatrics Society, 2003. 51(11): p. 1633-7.

42. Allali, G., et al., Frontotemporal dementia: pathology of gait? Movement Disorders,

2010. 25(6): p. 731-7.

43. Camicioli, R. and L. Licis, Motor impairment predicts falls in specialized Alzheimer care

units. Alzheimer Disease & Associated Disorders, 2004. 18(4): p. 214-8.

44. Gillain, S., et al., Gait speed or gait variability, which one to use as a marker of risk to

develop Alzheimer disease? A pilot study. Aging-Clinical & Experimental Research,

2016. 28(2): p. 249-255.

45. Hsu, Y.L., et al., Gait and balance analysis for patients with Alzheimer's disease using

an inertial-sensor-based wearable instrument. IEEE Journal of Biomedical & Health

Informatics, 2014. 18(6): p. 1822-30.

46. Lamoth, C.J., et al., Gait stability and variability measures show effects of impaired

cognition and dual tasking in frail people. Journal of Neuroengineering & Rehabilitation,

2011. 8: p. 2.

47. Nascimbeni, A., et al., Dual task-related gait changes in patients with mild cognitive

impairment. Functional Neurology, 2015. 30(1): p. 59-65.

48. Taylor, M.E., et al., Gait parameter risk factors for falls under simple and dual task

conditions in cognitively impaired older people. Gait & Posture, 2013. 37(1): p. 126-30.

49. Wittwer, J.E., K.E. Webster, and K. Hill, The effects of a concurrent motor task on

walking in Alzheimer's disease. Gait & Posture, 2014. 39(1): p. 291-6.

50. Allali, G., et al., Impact of impaired executive function on gait stability. Dementia &

Geriatric Cognitive Disorders, 2008. 26(4): p. 364-9.

-52-

51. Allali, G., et al., Changes in gait while backward counting in demented older adults with

frontal lobe dysfunction. Gait & Posture, 2007. 26(4): p. 572-6.

52. Beauchet, O., et al., Test-retest reliability of stride time variability while dual tasking in

healthy and demented adults with frontotemporal degeneration. Journal of

NeuroEngineering and Rehabilitation, 2011. 8(1).

53. Lin, Y.C., et al., Comparison of motor performance of upper and lower extremities in

dual-task tests in patients with mild Alzheimer's dementia. Aging Clinical and

Experimental Research, 2016. 28(3): p. 491-496.

54. König, A., et al., Objective measurement of gait parameters in healthy and cognitively

impaired elderly using the dual-task paradigm. Aging Clinical & Experimental

Research, 2017. 29(6): p. 1181-1189.

55. Winblad, B., et al., Mild cognitive impairment – beyond controversies, towards a

consensus: report of the International Working Group on Mild Cognitive Impairment.

Journal of Internal Medicine, 2004. 256(3): p. 240-246.

56. McKhann, G., et al., Clinical diagnosis of Alzheimer's disease Report of the NINCDS‐

ADRDA Work Group* under the auspices of Department of Health and Human Services

Task Force on Alzheimer's Disease. Neurology, 1984. 34(7): p. 939-939.

57. Lezak, M.D., et al., Neuropsychological assessment. 2004: Oxford University Press,

USA.

58. D'Esposito, M., et al., Maintenance versus Manipulation of Information Held in Working

Memory: An Event-Related fMRI Study. Brain and Cognition, 1999. 41(1): p. 66-86.

59. St Clair-Thompson, H.L. and R.J. Allen, Are forward and backward recall the same? A

dual-task study of digit recall. Memory & Cognition, 2013. 41(4): p. 519-532.

60. Rabin, L.A., et al., Differential memory test sensitivity for diagnosing amnestic mild

cognitive impairment and predicting conversion to Alzheimer's disease.

Neuropsychology, development, and cognition. Section B, Aging, neuropsychology and

cognition, 2009. 16(3): p. 357-376.

-53-

61. Yesavage, J.A., et al., Development and validation of a geriatric depression screening

scale: a preliminary report. Journal of Psychiatric Research, 1982. 17(1): p. 37-49.

62. Bjelland, I., et al., The validity of the Hospital Anxiety and Depression Scale: an updated

literature review. Journal of Psychosomatic Research, 2002. 52(2): p. 69-77.

63. Hausdorff, J.M., et al., Dual-Task Decrements in Gait: Contributing Factors Among

Healthy Older Adults. Journals of Gerontology Series a-Biological Sciences and Medical

Sciences, 2008. 63(12): p. 1335-1343.

64. Hausdorff, J.M., J. Balash, and N. Giladi, Effects of Cognitive Challenge on Gait

Variability in Patients with Parkinson’s Disease. Journal of Geriatric Psychiatry and

Neurology, 2003. 16(1): p. 53-58.

65. Hamilton, F., et al., Walking and talking: an investigation of cognitive—motor dual

tasking in multiple sclerosis. Multiple Sclerosis Journal, 2009. 15(10): p. 1215-1227.

66. Beauchet, O., et al., Association of increased gait variability while dual tasking and

cognitive decline: results from a prospective longitudinal cohort pilot study.

GeroScience, 2017. 39(4): p. 439-445.

67. Beauchet, O., et al., Stops walking when talking: A predictor of falls in older adults?

European Journal of Neurology, 2009. 16(7): p. 786-795.

68. Herssens, N., et al., Do spatiotemporal parameters and gait variability differ across the

lifespan of healthy adults? A systematic review. Gait & Posture, 2018. 64: p. 181-190.

69. Beauchet, O., et al., Poor Gait Performance and Prediction of Dementia: Results From

a Meta-Analysis. Journal of the American Medical Directors Association, 2016. 17(6):

p. 482-490.

70. Pettersson, A., E. Olsson, and L.-O. Wahlund, Motor function in subjects with mild

cognitive impairment and early Alzheimer’s disease. Dementia and Geriatric Cognitive

Disorders, 2005. 19(5-6): p. 299-304.

-54-

71. Allali, G., et al., Gait phenotype from mild cognitive impairment to moderate dementia:

results from the GOOD initiative. European Journal of Neurology, 2016. 23(3): p. 527-

541.

72. Nasreddine, Z.S., et al., The Montreal Cognitive Assessment, MoCA: A Brief Screening

Tool For Mild Cognitive Impairment. Journal of the American Geriatrics Society, 2005.

53(4): p. 695-699.

73. Kelly, V.E., A.J. Eusterbrock, and A. Shumway-Cook, A Review of Dual-Task Walking

Deficits in People with Parkinson's Disease: Motor and Cognitive Contributions,

Mechanisms, and Clinical Implications. Parkinson's Disease, 2012. 2012(1): p. 14.

74. Rhea, C.K., et al., Noise and Complexity in Human Postural Control: Interpreting the

Different Estimations of Entropy. PLoS ONE [Electronic Resource], 2011. 6(3): p.

e17696.

75. Wittwer, J.E., K.E. Webster, and K. Hill, The effects of a concurrent motor task on

walking in Alzheimer's disease. Gait & Posture, 2014. 39(1): p. 291-296.

-55-

FIGURE LEGENDS

FIGURE 2.1 Flow diagram of the systematic review process.

FIGURE 2.2 Forest plot for within study comparison for single-task and dual-task for all gait

dynamic outcomes and all dual-task procedures

FIGURE 2.3 Forest plot for within study comparison for cognitively impaired and cognitively

healthy for all gait dynamic outcomes and all dual-task procedures

FIGURE 2.4 Forest plot for meta-analysis of within study comparison for MCI and dementia

for all gait dynamic outcomes and all dual-task conditions.

-56-

FIGURE 2.1 Flow diagram of the systematic review process

Scre

enin

g In

clud

ed

Elig

ibili

ty

Iden

tific

atio

n Additional records identified through other sources

(n = 5)

Search results (n = 16,524)

Records excluded bases on title or

abstract (n = 9,543)

Full-text articles excluded, with reasons

(n = 24) (n = 15 Not full text articles; n = 4 No gait variability measure; n = 1 No MCI/dementia diagnosis; n = 3 Complete data not extractable; n = 1 Same data set as an included study)

Studies included in qualitative synthesis

(n = 25)

Duplicates removed (n = 6,932)

Records identified through database searching

(n = 16,519)

Records after duplicates removed

(n = 9,592)

Full-text articles assessed for eligibility

(n = 49)

-57-

FIGURE 2.2 Forest plot for within study comparison for single-task and dual-task for all gait

dynamic outcomes and all dual-task procedures

-58-

Forest plot indicates that dual-task gait was more varied than single-task gait in cognitively

impaired adults, which reflects worse gait dynamics under dual-task conditions.

MCI=Mild Cognitive Impairment; SD=Standard deviation; Std=Standardized; CI=Confidence

interval; CV=Coefficient of variation; I2=Measures heterogeneity.

-59-

FIGURE 2.3 Forest plot for within study comparison for cognitively impaired and cognitively

healthy for all gait dynamic outcomes and all dual-task procedures

Forest plot indicates that dual-task cost was larger in cognitively impaired older adults than

cognitively healthy adults, which reflects worse gait dynamics.

MCI=Mild Cognitive Impairment; SD=Standard deviation; Std=Standardized; CI=Confidence

interval; CV=Coefficient of variation; I2=Measures heterogeneity.

-60-

FIGURE 2.4 Forest plot for meta-analysis of within study comparison for MCI and dementia for all gait dynamic outcomes and all dual-task

conditions

Forest plot indicates that dual-task cost was larger in dementia than MCI, which reflects worse gait dynamics.

MCI=Mild Cognitive Impairment; SD=Standard deviation; Std.=Standardized; CI=Confidence interval; CV=Coefficient of variation;

I2=Measures heterogeneity.

-61-

TABLE 2.1 Medline full electronic search strategy example

No. Searches

1 'dual task*' or 'dual-task*' or 'multi task*' or 'multi-task' or 'secondary task*' or

'attention task*' or 'cognitive task*' or 'motor task*' or 'two task*' or '2 task*').mp.

2 walk* or gait or locomot* or ambulat* or stride* or step* or 'double limb' or 'double-

limb' or 'double support' or 'swing time' or 'stride-to-stride' or 'stride to stride' or 'foot

clearance' or mobility or stability or instability or 'gait variability' or 'centre of

pressure' or 'center of pressure' or COP or 'centre of mass' or 'center of mass' or

COM or ataxia or McRoberts or 'Gait Up' or APDM or GAITRite or Axivity or AX3

or Opal* or Pedar or Zeno or Gyroscope* or Lyap* or fractal).mp.

3 1 and 2

No.=number; mp=title, abstract, original title, name of substance word, subject heading word,

floating sub-heading word, keyword heading word, protocol supplementary concept word, rare

disease supplementary concept word, unique identifier, synonyms.

-62-

TABLE 2.2 Cohort characteristics Author, Year

[reference]

Participants

Pathology

n Gender (%female)

Age

(Yr)

ST gait speed

(m/s)

MMSE

Sheridan,

2003

A:AD A:28 NR A:77.9±6.9 A:0.57±0.20 A:13.8±7.9

Camicioli,

2004

A:AD (Non-faller)

B:AD (Faller)

A:24

B:18

A:91.7

B:77.8

A:82.3±6.7

B:83.1±9.6

A:0.70±0.17

B:0.62±0.25

A:14.7±7.2

B:15.8±7.6

Camicioli,

2006

A:AD (w/≤3 EPS)

B:AD (w/>3 EPS)

A:13

B:29

A:100.0

B:79.3

A:80.5±6.5

B:83.6±8.5

A:0.79±0.15

B:0.62±0.18

A:14.4±7.1

B:15.5±7.5

Allali, 2007 A:Dementia A:16 A:83.6 A:87.5 NR A:22.1±3.6

Allali, 2008 C:Control

A:AD

B:CI (w/IEF)

C:22

A:16

B:18

C:91.0

A:69.0

B:83.0

C:79.5 (8)

A:78.5 (8)

B:79.5 (5)

NR C:30.0 (1)

A:22.0 (4)

B:20.5 (6)

Gillian, 2009 C:Control

A:MCI

B:AD

C:14

A:14

B:6

C:50.0

A:50.0

B:50.0

C:73.5

A:72.9

B:73.7

C:1.4±0.13

A:1.22±0.15

B:1.02±0.36

C:28.2±1.6

A:26.7±1.7

B:22.8±2.1

Allali, 2010 C:Control

A:Dementia (bvFTD)

B:AD

C:22

A:19

B:19

C:63.6

A:47.4

B:68.4

C:71.0±0.5

A:66.8±9.7

B:79.3±8.4

C:1.19±11.7

A:1.12±9.0

B:1.11±9.9

C:29.0±1.0

A:26.0±6.0

B:19.0±7.0

-63-

Beauchet,

2011

C:Control

A:Dementia (w/FTD)

C:69

A:14

C:43.5

A:7.1

C:75.5±4.3

A:65.7±9.8

NR A:23.3±6.6

Lamoth, 2011 C:Control

A:AD

C:13

A:13

C:53.8

A:69.2

C:79.4±5.6

A:82.6±4.3

C:0.95±0.21

A:0.88±0.27

C:28.2±1.1

A:18.0±3.5

Ijmker, 2012 C:Control (older)

A:D

C:14

A:15

C:14.3

A:13.3

C:76.9±4.1

A:81.7±6.3

C:1.14±0.11

A:0.67±2.1

C:28.5±1.2

A:19.6±3.6

Montero-

Odasso, 2012

C:Control

A:MCI

C:25

A:43

C:88.0

A:54.0

C:71.5±4.1

A:75.1±6.3

NR C:29.5±0.6

A:27.8±1.6

Muir, 2012 C:Control

A:MCI

B:AD

C:22

A:29

B:23

C:88.0

A:59.0

B:61.0

C:71.0±5.0

A:73.6±6.2

B:77.5±5.0

C:1.36±0.24

A:1.16±0.21

B:1.11±0.14

C:29.5±0.6

A:27.5±1.9

B:24.2±2.3

Taylor, 2013 A:CI (Non-multiple

fallers)

B:CI (Multiple fallers)

A:41

B:22

A:46.3

B:45.5

A:80.7±6.8

B:82.5±6.9

A:0.94±0.24

B:0.79±0.30

A:24.8±3.6

B:22.7±5.1

Beauchet,

2014

A:AD A:86 A:68.6 A:82.6±5.5 A:0.63±0.21 A:17.6±5.5

Hsu, 2014 C:Control

A:AD

C:50

A:21

C:40.0

A:52.4

C:59.9±4.6

A:61.5±4.9

C:1.38±0.17

A:1.25±0.16

C:28.4±1.6

A:23.0±3.2

-64-

Montero-

Odasso, 2014

A:MCI (Na)

B:MCI (a)

A:22

B:42

A:63.6

B:42.9

A:74.2±6.5

B:77.3±7.3

A:1.09±0.19

B:1.00±0.22

A:29.1±0.8

B:27.2±2.1

Wittwer,

2014

A:AD A:30 A:50.0 A:80.2±5.8 A:1.12±0.27 A:20.6±5.1

Nascimbeni,

2015

C:Control

A:MCI

C:10

A:13

C:40.0

A:15.4

C:72.0±3.9

A:76.0±3.9

C:0.97±0.15

A:0.83±0.21

NR

Gillian, 2016 A:MCI (future AD)

B:MCI

A:9

B:4

A:44.4

B:50.0

A:74.4±4.16

B:70.0±2.16

A:1.15±0.13

B:1.29±0.10

A:26.1±1.5

B:27.3±1.7

Lin, 2016 C:Control

A:AD

C:10

A:10

C:80.0

A:80.0

C:73.8±6.1

A:74.0±8.6

NR C:29.4±0.7

A:17.7±4.1

Martinez-

Ramirez,

2016

C:Control (frail)

A:MCI (frail)

C:20

A:11

C:70.0

A:72.7

C:93.4±3.2

A:92.4±4.2

C:0.68±0.26

A:0.77±0.18

NR

Auvinet, 2017 A:MCI A:24 A:33.3 A:76.4±5.8 A:1.0±0.3 A:26.2±2.1

Gschwind,

2017

A:MCI A:50 A:50.0 A:68.5±8.4 A:1.27±0.18 NR

Konig, 2017 C:Control

A:AD

B:MCI

C:22

A:23

B:24

C:68.2

A:47.8

B:66.7

C:73.0±7.0

A:77.0±9.0

B:75.0±9.0

NR C:28.4±1.5

A:17.0±4.6

B:24.8±3.2

-65-

Lee, 2018 C:Control

A:MCI

C:8

A:8

NR C:66.1±1.6

A:66.5±1.9

C:1.06±0.06

A:1.00±0.18

C:28.1±0.8

A:21.0±0.8

Results are presented as Mean±Standard Deviation or Median (interquartile range); n=number;

%=percent; Yr=years; kg=kilograms; m/s=meters per second; C=control; A/B=intervention

group A or B; w/=with; IEF=impaired executive function; D=dementia; AD=Alzheimer's

disease; FTD=Frontotemporal degeneration; EPS=extra-pyramidal signs; MCI=mild cognitive

impairment; CI=cognitive impairment; bvFTD=behavioral variant of frontotemporal

degeneration; NR=not reported; ST=Single-task.

-66-

TABLE 2.3 Risk of bias assessment Author, Year [reference]

Reporting, item no. External validity, item no.

Internal validity, item no. Total /24 Bias Confounding

1 2 3 4 5 6 7 8 10 11 12 13 14 15 16 18 19 20 21 22 23 24 25 27 Sheridan, 2003 1 1 1 1 0 1 1 0 1 0 0 1 0 0 1 1 1 1 1 0 0 0 0 0 13 Camicioli, 2004 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 0 0 1 0 14 Camicioli, 2006 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 0 0 1 0 14 Allali, 2007 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 1 0 0 0 12 Allali, 2008 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 1 0 1 0 13 Gillian, 2009 1 1 1 1 1 1 1 0 0 0 0 1 0 0 1 1 1 1 1 0 0 0 0 0 13 Allali, 2010 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 1 0 1 0 15 Beauchet, 2011 1 1 1 1 0 1 1 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 11 Lamoth, 2011 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 1 13 Ijmker, 2012 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 1 12 Montero-Odasso, 2012 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 1 0 1 1 16 Muir, 2012 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 0 0 1 0 0 0 14 Taylor, 2013 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 12 Beauchet, 2014 1 1 1 1 1 1 1 0 1 1 0 1 0 0 1 1 1 1 1 1 0 0 0 1 17 Hsu, 2014 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 12 Montero-Odasso, 2014 1 1 1 1 1 1 1 0 1 0 0 0 0 0 1 1 1 1 1 0 1 0 1 1 16 Wittwer, 2014 1 1 1 1 0 1 1 1 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 13 Nascimbeni, 2015 1 1 1 1 1 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 1 0 1 1 15 Gillian, 2016 1 1 1 1 0 1 1 0 1 0 0 1 0 0 1 1 1 1 1 1 0 0 0 0 14 Lin, 2016 1 0 0 1 1 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 1 0 0 0 11 Martinez-Ramirez, 2016 1 1 1 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 1 12 Auvinet, 2017 1 1 1 1 1 1 1 0 1 1 0 1 0 0 1 1 1 1 1 0 0 0 1 0 16 Gschwind, 2017 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1 1 1 1 1 1 1 0 0 1 17

-67-

Konig, 2017 1 1 1 1 1 1 1 0 1 0 0 1 0 0 1 1 1 1 1 0 0 0 1 0 15 Lee, 2018 1 1 0 1 0 1 1 0 1 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 10

Results are presented for each criterion with a total score presented under the ‘Total’ column with a maximum score of 24. The criteria that were

not included for scoring were: items 9, 17 and 26 (relating to follow up). Items modified to be consistent with the scoring procedure included:

item number 5 (relating to confounder distribution), and item number 27 (relating to statistical power), i.e., alteration from a 0 to 2 scale, and a 0

to 5 scale, to a 'no, 0; unable to determine, 0; and yes, 1'. Criterion 23 was altered for increased specificity to read: ‘Was the order of the walking

tasks (single and dual) randomized for study subjects?’.

-68-

TABLE 2.4 Dual-task procedure characteristics

Author, Year [reference]

Measurement system

Dual-task type Dual -task condition(s)

Measured distance walked (m)

Number of trials per condition (n)

Order of walking randomized

Sheridan, 2003

Footswitch sensors

Working memory Forward digit span 15.24-152.4 (up to 10 laps) Avg distance 106.68 (7laps)

NR Set order

Camicioli, 2004

GAITRite Working memory Forward counting from 1 by 1s

3.66 ST: 2 (used 2nd trial data only), DT: 1

Set order

Camicioli, 2006

GAITRite Working memory Forward counting from 1 by 1s

3.66 ST: 2 DT: 1

Set order

Allali, 2007 GAITRite Working memory Mental tracking

Forwards counting Backwards counting

10 1 Randomized

Allali, 2008 GAITRite Mental tracking Backwards counting from 50 by 1s

10 NR Randomized

Gillian, 2009 Locometrix Mental tracking Backwards counting from 50 by 1s

30 3 (Used 2nd trial data only)

NR

Allali, 2010 SMTEC footswitch

Mental tracking Backwards counting from 50 by 1s

10 1 Randomized

Beauchet, 2011

GAITRite & SMTEC footswitch system

Mental tracking Backwards counting from 50 by 1s

3.5 (GAITRite) & 10 (SMTEC)

2 NR

-69-

Lamoth, 2011 DynaPort MiniMod

Verbal fluency Name words starting with either 'R' or 'G'

~160m (3 min in 40m corridor)

1 NR

Ijmker, 2012 DynaPort MiniMod

Verbal fluency Name animals, occupations or words starting with either 'R', 'G' or 'P' in 1 minute.

10 1 NR

Montero-Odasso, 2012

GAITRite Mental tracking Verbal fluency

Subtracting serial 7s from 100 Naming animals

6 NR (1st trial as a practice)

Randomized

Muir, 2012 GAITRite Mental tracking Verbal fluency Mental tracking

Backwards counting from 100 by 1s Naming animals Subtracting serial 7s from 100

6 NR (1st trial as a practice)

Randomized

Taylor, 2013 GAITRite Motor Mental tracking

Carrying a glass filled (10mm from rim) of water Backwards counting from 30 by 1s

4.6 2 NR

Beauchet, 2014

GAITRite Mental tracking Backwards counting from 50 by 1s

7.92 1 NR

Hsu, 2014 Wearable device Mental tracking Backwards counting from 100 by 1s

40 NR NR

-70-

Montero-Odasso, 2014

GAITRite Mental tracking Mental tracking Verbal fluency

Backwards counting from 100 by 1s Subtracting serial 7s from 100 Naming animals

6 NR (1st trial as a practice)

Randomized

Wittwer, 2014 GAITRite Motor Carrying a tray with two empty glasses using both hands

8.3 2 to 4 (mean of trials used)

Set order

Nascimbeni, 2015

STEP 32 system & 3 footswitch sensors

Verbal fluency Verbal memory Mental tracking

Naming words beginning with F, A or S for 1 minute Short story recall Backwards counting from either 378 or 283 by 1s

12 1 Randomized

Gillian, 2016 Locometrix Mental tracking Backwards counting from 50 by 1s

30 1 Set order

Lin, 2016 Vicon MX infrared camera and 3 AMTI force plates

Working memory Mental tracking

Forward 3-digit working task Backwards 3-digit working task

8 6 (1st trial as a practice)

Randomized

Martinez-Ramirez, 2016

Orientation Tracker MTx

Mental tracking Verbal fluency

Backwards counting from 100 by 1s Naming animals

5 NR NR

-71-

Auvinet, 2017 Locometrix Mental tracking Backwards counting from 50 by 1s

30 1 Set order

Gschwind, 2017

GAITRite Mental tracking Verbal fluency

Backwards counting from 50 by 2s Naming animals

10 1 Randomized

Konig, 2017 CE-marked accelerometer

Mental tracking Backwards counting from 305 by 1s.

~20 (10 up and 10 back plus turn)

1 NR

Lee, 2018 3D movement analysis

Mental tracking Backwards counting from 100 by 1s

12 1 Set order

m=meters; n=number; ST=Single-task; DT=Dual-task; NR=not reported; all verbal fluency tasks that used a word naming task using a specific

letter stated that the letter choice was predetermined; further detail for the task in Sheridan, 2003 was to repeat a list of random single-digit numbers

forward starting with 2 digits and progressing to 8 digits.

-72-

TABLE 2.5a Gait outcomes for Mild Cognitive Impairment group: single-task vs. dual-task comparisons Author, Year [reference]

Pathology Task Single-task Mean±SD

Dual-task Mean±SD

Mean Difference (95% CI)

Between group ES (95% CI)

Stride time CV

Montero-Odasso, 2012

MCI Backwards counting from 100 by 7s

2.68±1.31 9.84±10.13 7.16 [6.38, 7.94] 5.37 [4.03, 6.70]

MCI Animal naming 2.68±1.31 7.16±7.76 4.48 [3.70, 5.26] 3.36 [2.40, 4.31]

Muir, 2012 MCI Backwards counting from 100 by 1s

2.59±1.47 4.06±2.37 1.47 [0.40, 2.54] 0.97 [0.20, 1.75]

MCI Animal naming 2.59±1.47 8.02±8.88 5.43 [4.36, 6.50] 3.59 [2.36, 4.82]

MCI Backwards counting

from 100 by 7s 2.59±1.47 10.07±9.29 7.48 [6.41, 8.55] 4.95 [3.39, 6.50]

Montero-Odasso, 14

MCI (na) Backwards counting from 100 by 1s

2.40±1.38 2.90±0.98 0.50 [-0.65, 1.65] 0.35 [-0.49, 1.19]

MCI (a) Backwards counting from 100 by 1s

3.33±2.60 4.81±3.73 1.51 [-0.06, 3.08] 0.57 [-0.05, 1.19]

MCI (na) Backwards counting from 100 by 7s

2.40±1.38 3.82±2.10 1.42 [0.27, 2.57] 0.99 [0.09, 1.89]

MCI (a) Backwards counting from 100 by 7s

3.33±2.60 5.63±5.00 2.30 [0.73, 3.87] 0.87 [0.23, 1.50]

-73-

MCI (na) Animal naming 2.40±1.38 4.83±3.53 2.43 [1.28, 3.58] 1.69 [0.69, 2.70]

MCI (a) Animal naming 3.33±2.60 6.47±5.71 3.14 [1.57, 4.71] 1.18 [0.52, 1.85]

Nascimbeni, 2015 MCI Word naming with 'F', 'A', or 'S'

3.17±1.12 5.52±3.04 2.35 [1.13, 3.57] 1.95 [0.54, 3.36]

MCI Short story recall 3.17±1.12 5.42±1.81 2.25 [1.03, 3.47] 1.87 [0.48, 3.26]

MCI Backwards counting

from 378 or 283 by 1s 3.17±1.12 5.07±3.30 1.90 [0.68, 3.12] 1.58 [0.27, 2.89]

Gschwind, 2017

MCI Backwards counting from 50 by 12s

1.95±0.90 6.40±16.20 4.45 [3.95, 4.95] 4.87 [3.73, 6.00]

MCI Animal naming 1.95±0.90 8.65±16.10 6.70 [6.20, 7.20] 7.33 [5.73, 8.92]

Lee, 2018 MCI Backwards counting from 100 by 1s

2.44±0.99 11.21±6.84 8.77 [7.40, 10.14] 7.70 [2.23, 13.18]

Step regularity

Gillian, 2009 MCI Backwards counting from 50 by 1s

287.00±29.00 224.00±47.00 -63.00 [-93.38, -32.62] 2.03 [0.66, 3.41]*

Martinez-Ramirez, 2016

MCI (frail) Backwards counting from 100 by 1s

0.58±0.18 0.51±0.21 -0.07 [-0.28, 0.14] 0.36 [-0. 85, 1.56]*

MCI (frail) Naming animals 0.58±0.18 0.37±0.21 -0.21 [-0.42, 0.00] 1.07 [-0.24, 2.38]*

-74-

Stride regularity

Gillian, 2016

MCI (future AD) Backwards counting from 50 by 1s

286.20±37.45 220.67±254.88 -65.53 [-114.77, -16.29]

1.56 [-0.07, 3.18]*

MCI (-) Backwards counting from 50 by 1s

298.00±22.46 254.88±32.86 -43.12 [-87.14, 0.09] 1.10 [-5.41, 5.61]*

Martinez-Ramirez, 2016

MCI (frail) Backwards counting from 100 by 1s

0.59±0.16 0.44±0.19 -0.15 [-0.34, 0.04] 0.86 [-0.41, 2.13]*

MCI (frail) Naming animals 0.59±0.16 0.38±0.14 -0.21 [-0.40, -0.02] 1.20 [-0.14, 2.54]*

Auvinet, 2017 MCI Backwards counting from 50 by 1s

214.70±54.20 156.60±65.20 -58.10 [-14.73, -101.47]

1.03 [0.17, 1.90]*

Step time CV

Martinez-Ramirez, 2016

MCI (frail) Backwards counting from 100 by 1s

0.09±0.02 0.11±0.05 0.02 [0.00, 0.04] 0.91 [-0.36, 2.19]

MCI (frail) Naming animals 0.09±0.02 0.15±0.06 0.06 [0.04, 0.08] 2.74 [0.88, 4.60]

Step time variance

Konig, 2017 MCI Backwards counting from 305 by 1s

5.70±4.50 6.80±5.30 1.10 [-2.50, 4.70] 0.24 [-0.57, 1.04]

Approximate entropy

Martinez-Ramirez, 2016

MCI (frail) Backwards counting from 100 by 1s

0.16±0.12 2.12±1.65 1.96 [1.82, 2.10] 14.93 [7.05, 22.81]

MCI (frail) Naming animals 0.16±0.12 3.7±2.41 3.54 [3.40, 3.68] 26.97 [12.85,

41.09]

-75-

SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; MCI=mild cognitive impairment; (a)=amnestic;

(na)=non-amnestic; *=ES was reversed to indicate change in the direction of other outcome measures. ESs were calculated as standardized mean

difference and 95% CI.

-76-

TABLE 2.5b Gait outcomes for dementia (including Alzheimer’s disease) group: single-task vs. dual-task comparisons

Author, Year [reference]

Pathology Task Single-task Mean±SD

Dual-task Mean±SD

Mean Difference (95% CI)

Between group ES (95% CI)

Stride time CV Sheridan, 2003 AD Forward digit

span 8.50±3.40 11.10±5.50 2.60 [0.08, 5.12] 0.74 [-0.03, 1.51]

Camicioli, 2006

AD (w/≤3 EPS)

Forwards counting from 1 by 1s

4.04±1.94 4.65±2.83 0.61 [-1.51, 2.71] 0.29 [-0.81, 1.39]

AD (w/>3EPS) Forwards

counting from 1 by 1s

4.36±3.81 7.78±12.20 3.42 [0.65, 6.19] 0.87 [0.11, 1.64]

Allali, 2007 Cognitive impairment

Forwards counting

4.00±2.20 7.60±10.00 3.60 [1.44, 5.76] 1.55 [0.39, 2.71]

Cognitive impairment

Backwards counting

4.00±2.20 15.40±16.10 11.40 [9.24, 13.56] 4.90 [2.71, 7.15]

Allali, 2008

AD Backwards counting from 50 by 1s

1.17±0.56 4.10±3.47 2.93 [2.38, 3.48] 4.95 [2.74, 7.15]

Cognitive impairment (w/IEF)

Backwards counting from 50 by 1s

2.77±2.02 14.33±13.72 11.56 [9.69, 13.43] 5.45 [3.23, 7.67]

-77-

Allali, 2010

bvFTD Backwards counting from 50 by 1s

7.70±8.20 8.30±6.20 0.60 [-6.78, 7.98] 0.07 [-0.83, 0.97]

AD Backwards counting from 50 by 1s

3.10±1.20 6.00±3.10 2.90 [1.82, 3.98] 2.31 [1.09, 3.53]

Beauchet, 2011 Dementia (w/IEF)

Backwards counting from 50 by 1s

5.00±2.50 7.60±6.70 2.60 [-0.02, 5.22] 0.97 [-0.16, 2.10]

Lamoth, 2011 AD Word naming with 'R' or 'G'

4.20±2.7 3.67±1.67 -0.53 [-3.47, 2.41] -0.18 [-1.28, 0.91]

Ijmker, 2012 AD Word naming with 'R', 'G' or 'P'

9.88±5.28 12.88±6.78 3.00 [-2.36, 8.36] 0.53 [-0.50, 1.57]

Muir, 2012 AD Backwards counting from 100 by 1s

2.67±1.08 4.86±2.74 2.19 [1.31, 3.07] 1.95 [0.93, 2.98]

AD Animal naming 2.67±1.08 9.04±8.94 6.37 [5.49, 7.25] 5.68 [3.70, 7.67]

AD Backwards

counting from 100 by 7s

2.67±1.08 12.49±12.33 9.82 [8.94, 10.70] 8.76 [5.86, 11.66]

Beauchet, 2014

AD Backwards counting from 50 by 1s

5.61±4.00 8.85±7.40 3.24 [1.55, 4.93] 0.80 [0.36, 1.24]

Hsu, 2014 AD Backwards counting from 100 by 1s

2.31±0.66 31.31±26.83 29.00 [28.43, 29.57] 42.18 [28.00, 56.36]

-78-

Wittwer, 2014 AD Carrying a tray with two empty glasses using both hands

2.40±0.80 2.80±0.80 0.40 [-0.17, 0.97] 0.49 [-0.24, 1.21]

Lin, 2016 AD Forwards 3-digit span

5.20±1.90 5.80±5.00 0.60 [-1.76, 2.96] 0.29 [-0.96, 1.54]

AD Backwards 3-

digit span 5.20±1.90 9.90±3.80 4.70 [2.34, 7.06] 2.23 [0.47, 4.00]

Stride length CV

Camicioli, 2006

AD (w/≤3 EPS)

Forwards counting from 1 by 1s

5.73±3.36 4.27±2.33 -1.46 [-2.34, 5.26] -0.40 [-0.75, 1.55]

AD (w/>3 EPS)

Forwards counting from 1 by 1s

5.78±2.25 5.59±2.51 -0.19 [-1.48, 1.86] -0.08 [-0.66, 0.82]

Taylor, 2013

Cognitive impairment (non-multiple faller)

Carrying a glass filled with water

2.63±1.62 3.17±2.05 0.54 [-0.46, 1.54] 0.33 [-0.30, 0.95]

Cognitive impairment (multiple faller)

Carrying a glass filled with water

4.20±2.54 5.76±4.66 1.56 [-0.56, 3.68] 0.59 [-0.27, 1.45]

Cognitive impairment (non-multiple faller)

Backwards counting from 30 by 1s

2.63±1.62 3.74±2.83 1.11 [0.11, 2.11] 0.67 [0.03, 1.31]

-79-

Cognitive impairment (multiple faller)

Backwards counting from 30 by 1s

4.20±2.54 6.34±5.03 2.14 [0.02, 4.26] 0.81 [-0.07, 1.69]

Wittwer, 2014 AD Carrying a tray with two empty glasses using both hands

3.20±1.00 3.90±1.50 0.70 [-0.02, 1.42] 0.68 [-0.06, 1.42]

Lin, 2016 AD Forwards 3-digit span

6.70±5.30 7.70±3.40 1.00 [-5.57, 7.57] 0.17 [-1.07, 1.41]

AD Backwards 3-

digit span 6.70±5.30 10.40±2.30 3.70 [-2.87, 10.27] 0.63 [-0.66, 1.92]

Swing time CV

Camicioli, 2006

AD (w/≤3 EPS)

Forwards counting from 1 by 1s

6.15±4.19 8.58±4.53 2.43 [-2.14, 7.00] 0.54 [-0.58, 1.66]

AD (w/>3 EPS)

Forwards counting from 1 by 1s

7.95±3.92 11.09±7.40 3.14 [0.28, 6.00] 0.78 [0.02, 1.54]

Taylor, 2013

Cognitive impairment (non-multiple faller)

Carrying a glass filled with water

6.25±3.39 6.74±5.30 0.49 [-1.59, 2.57] 0.14 [-0.47, 0.75]

Cognitive impairment (multiple faller)

Carrying a glass filled with water

8.07±4.37 10.65±7.67 2.58 [-1.07, 6.23] 0.57 [-0.29, 1.42]

-80-

Cognitive impairment (non-multiple faller)

Backwards counting from 30 by 1s

6.25±3.39 9.42±7.31 3.17 [1.09, 5.25] 0.92 [0.27, 1.56]

Cognitive impairment (multiple faller)

Backwards counting from 30 by 1s

8.07±4.37 15.06±11.25 6.99 [3.34, 10.64] 1.54 [0.56, 2.51]

Hsu, 2014 AD Backwards counting from 100 by 1s

2.67±0.69 17.24±12.99 14.57 [13.98, 15.16] 20.27 [13.42, 27.13]

Walking speed CV

Beauchet, 2014

AD Backwards counting from 50 by 1s

8.25±5.00 11.49±7.10 3.24 [1.13, 5.35] 0.64 [0.21, 1.08]

Base of support CV

Camicioli, 2006

AD (w/≤3 EPS)

Forwards counting from 1 by 1s

23.7±13.45 29.6±22.80 5.90 [-8.77, 20.57] 0.41 [-0.70, 1.54]

AD (w/>3 EPS)

Forwards counting from 1 by 1s

18.7±15.3 19.3±11.90 0.60 [-10.54, 11.74] 0.04 [-0.69, 0.77]

Double support time CV

Camicioli, 2006

AD (w/≤3 EPS)

Forwards counting from 1 by 1s

11.20±6.96 8.65±4.88 -2.55 [-10.14, 5.04] -0.34 [-0.44, 0.76]

-81-

AD (w/>3 EPS)

Forwards counting from 1 by 1s

15.10±11.60 19.70±20.3 4.60 [-3.85, 13.05] 0.39 [-0.35, 1.12]

Left stride length CV

Camicioli, 2004

AD (non-faller) Forwards counting from 1 by 1s

5.79±2.95 5.25±3.47 -0.54 [-2.90, 1.82] -0.18 [-0.98, 0.63]

AD (faller) Forwards counting from 1 by 1s

5.89±3.20 5.67±3.48 -0.22 [-3.18, 2.74] -0.07 [-0.99, 0.86]

Left base of support CV

Camicioli, 2004

AD (non-faller) Forwards counting from 1 by 1s

15.13±10.82 24.71±21.95 9.58 [0.92, 18.24] 0.85 [0.01, 1.70]

AD (faller) Forwards counting from 1 by 1s

22.22±21.58 18.06±10.11 -4.16 [-24.10, 15.78] -0.18 [-1.11, 0.74]

Right stride length CV

Camicioli, 2004

AD (non-faller) Forwards counting from 1 by 1s

5.38±2.67 5.25±3.82 -0.13 [-2.27, 2.01] -0.05 [-0.85, 0.75]

AD (faller) Forwards counting from 1 by 1s

6.11±3.60 4.83±2.57 -1.28 [-4.61, 2.05] -0.34 [-1.27, 0.59]

Right base of support CV

Camicioli, 2004

AD (non-faller) Forwards counting from 1 by 1s

22.67±21.39 23.42±22.61 0.75 [-16.37, 17.87] 0.03 [-0.77, 0.83]

-82-

AD (faller) Forwards counting from 1 by 1s

22.06±21.16 22.72±21.61 0.66 [-18.89, 20.21] 0.03 [-0.89, 0.95]

Step regularity

Gillian, 2009 AD Backwards counting from 50 by 1s

227.00±82.00 139.00±81.00 -88.00 [-219.22, -43.22]

0.86 [-0.94, 2.66]*

Stance time CV

Hsu, 2014 AD Backwards counting from 100 by 1s

3.31±0.84 43.31±37.32 40.00 [39.28, 40.72] 45.71 [30.35, 61.80]

Stance period CV

Hsu, 2014 AD Backwards counting from 100 by 1s

4.78±5.48 12.3±10.18 7.52 [2.83, 12.21] 1.32 [0.35, 2.32]

Swing period CV

Hsu, 2014 AD Backwards counting from 100 by 1s

5.74±7.80 18.95±15.09 13.21 [6.53, 19.89] 1.63 [0.61, 2.64]

Step time variance

Konig, 2017 AD Backwards counting from 305 by 1s

6.70±7.10 10.20±9.90 3.50 [-2.31, 9.31] 0.47 [-0.36, 1.31]

Stride time DFA Lamoth, 2011 AD Word naming

with 'R' or 'G' 0.84±0.16 0.84±0.11 0.00 [-0.17, 0.17] 0.00 [-1.09, 1.09]

-83-

SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; AD=Alzheimer’s disease; w/=with; EPS=extra-

pyramidal signs; IEF=Impaired executive function; bvFTD=Behvioural variant frontotemporal degeneration; FTD=Frontotemporal degeneration;

*=ES was reversed to indicate change in the direction of other outcome measures. ESs were calculated as standardized mean difference and 95%

CI.

-84-

TABLE 2.6a Gait outcomes for cognitive status comparisons: Control vs. Mild Cognitive Impairment

Author, Year [reference] Task

Control group Cognitive impairment group

Mean Difference (95% CI)

Between group ES (95% CI) Single-task

Mean±SD Dual-task Mean±SD

Single-task Mean±SD

Dual-task Mean±SD

Stride time CV

Montero-Odasso, 2012

Backwards counting from 100 by 7s

1.86±0.66 3.74±3.31 2.68±1.31 9.84±10.13 5.28 [4.73, 5.83] 4.67 [3.72,5.61]

Naming animals 1.86±0.66 3.59±2.95 2.68±1.31 7.16±7.76 2.75 [2.20, 3.30] 2.43 [1.78, 3.08]

Muir, 2012 Backwards counting from 100 by 1s

1.72±0.66 2.12±1.35 2.59±1.47 4.06±2.37 1.07 [0.41, 1.73] 0.88 [0.30, 1.47]

Naming animals 1.72±0.66 2.69±1.57 2.59±1.47 8.02±8.88 4.46 [3.80, 5.12] 3.68 [2.76, 4.61]

Backwards counting from 100 by 7s

1.72±0.66 3.14±2.18 2.59±1.47 10.07±9.29 6.06 [5.40, 6.72] 5.00 [3.85, 6.16]

Nascimbeni, 2015

Word naming with 'F', 'A', or 'S'

3.58±1.99 4.44±1.69 3.17±1.12 5.52±3.04 1.49 [0.21, 2.77] 0.92 [0.05, 1.80]

-85-

Short story recall 3.58±1.99 5.57±2.60 3.17±1.12 5.42±1.81 0.26 [-1.02, 1.54] 0.16 [-0.66, 0.99]

Backwards counting from 378 or 283 by 1s

3.58±1.99 3.93±1.70 3.17±1.12 5.07±3.30 2.09 [0.81, 3.37] 1.30 [0.38, 2.22]

Lee, 2018 Backwards counting from 100 by 1s

2.77±1.05 4.49±1.90 2.44±0.99 11.21±6.84 7.05 [6.05, 8.05] 6.53 [3.75, 9.32]

Step regularity Gillian, 2009 Backwards counting

from 50 by 1s 276.00±35.00 258.00±38.00 287.00±29.00 224.00±47.00 -45.00 [-68.81, -21.19] 1.36 [0.52, 2.19]*

Martinez-Ramirez, 2016

Backwards counting from 100 by 1s

0.48±0.21 0.47±0.17 0.58±0.18 0.51±0.21 -0.06 [-0.21, 0.09] 0.29 [0.45, 1.03]*

Naming animals 0.48±0.21 0.40±0.21 0.58±0.18 0.37±0.21 -0.13 [-0.28, 0.02] 0.63 [0.12, 1.39]*

Stride regularity Martinez-Ramirez, 2016

Backwards counting from 100 by 1s

0.45±0.20 0.43±0.20 0.59±0.16 0.44±0.19 -0.13 [-0.27, 0.01] 0.68 [0.08, 1.43]*

Naming animals 0.45±0.20 0.36±0.22 0.59±0.16 0.38±0.14 -0.12 [-0.26, 0.02] 0.62 [0.13,1.38]*

Step time CV

-86-

Martinez-Ramirez, 2016

Backwards counting from 100 by 1s

0.12±0.05 0.13±0.06 0.09±0.02 0.11±0.05 0.01 [-0.02, 0.04] 0.23 [-0.51, 0.97]

Naming animals 0.12±0.05 0.14±0.07 0.09±0.02 0.15±0.06 0.04 [0.01, 0.07] 0.92 [0.15, 1.70]

Step time variance Konig, 2017 Backwards counting

from 305 by 1s 4.50±4.90 3.90±5.40 5.70±4.50 6.80±5.30 1.70 [-1.02, 4.42] 0.36 [-0.23, 0.94]

Approximate entropy

Martinez-Ramirez, 2016

Backwards counting from 100 by 1s

0.27±0.36 2.42±2.89 0.16±0.12 2.12±1.65 -0.19 [-0.41, 0.03] -0.62 [-1.37, 0.14]

Naming animals 0.27±0.36 3.67±4.12 0.16±0.12 3.7±2.41 0.14 [-0.08, 0.36] 0.45 [-0.29, 1.20]

SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; *=ES was reversed to indicate change in the

direction of other outcome measures. ESs were calculated as standardized mean difference and 95% CI.

-87-

TABLE 2.6b Gait outcomes for cognitive status comparisons: Control vs. dementia (including Alzheimer’s disease)

Author, Year [reference] Task

Control group Cognitive impairment group

Mean Difference (95% CI)

Between group ES (95% CI) Single-task

Mean±SD Dual-task Mean±SD

Single-task

Mean±SD

Dual-task Mean±SD

Stride time CV

Allali, 2008 Backwards counting from 50 by 1s

1.47±1.11 2.17±1.43 1.17±0.56 4.10±3.47 2.23 [1.64, 2.82] 2.37 [1.51, 3.22]

Backwards counting from 50 by 1s

1.47±1.11 2.17±1.43 2.77±2.02 14.33±13.72 10.86 [9.87, 11.85] 6.74 [5.06, 8.41]

Allali, 2010 Backwards counting from 50 by 1s

1.70±0.50 2.70±0.90 7.70±8.00 8.30±6.20 -0.40 [-3.83, 3.03] -0.07 [-0.68, 0.54]

Backwards counting from 50 by 1s

1.70±0.50 2.70±0.90 3.10±1.20 6.00±3.10 1.90 [1.35, 2.45] 2.09 [1.32, 2.87]

Beauchet, 2011 Backwards counting from 50 by 1s

1.30±1.00 1.70±1.40 5.00±2.50 7.60±6.70 2.20 [1.42, 2.98] 1.61 [0.98, 2.23]

Lamoth, 2011 Word naming with 'R' or 'G'

2.95±1.77 5.00±2.67 4.2±2.70 3.67±1.67 -2.58 [-4.33, -0.83] -1.09 [-1.93, -0.26]

Ijmker, 2012 Word naming with 'R', 'G' or 'P'

3.51±0.88 4.26±1.0 9.88±5.28 12.88±6.78 2.25 [-0.55, 5.05] 0.57 [-0.18, 1.31]

-88-

Muir, 2012 Backwards counting from 100 by 1s

1.72±0.66 2.12±1.35 2.67±1.08 4.86±2.74 1.79 [1.26, 2.32] 1.95 [1.23, 2.68]

Animal naming 1.72±0.66 2.69±1.57 2.67±1.08 9.04±8.94 5.40 [4.87, 5.93] 5.90 [4.49, 7.30]

Backwards counting from 100 by 7s

1.72±0.66 3.14±2.18 2.67±1.08 12.49±12.33 8.40 [7.87, 8.93] 9.17 [7.10, 11.24]

Hsu, 2014 Backwards counting from 100 by 1s

2.02±0.78 14.38±18.37

2.31±0.66 31.31±26.83 16.64 [16.26, 17.02] 22.03 [18.26, 25.79]

Lin, 2016 Forwards 3-digit span

3.40±1.90 4.30±1.20 5.20±1.90 5.80±5.00 -0.30 [-1.97, 1.37] -0.15 [-1.03, 0.73]

Backwards 3-digit span

3.40±1.90 4.20±1.80 5.20±1.90 9.90±3.80 3.90 [2.23, 5.57] 1.97 [0.86, 3.08]

Stride length CV Lin, 2016 Forwards 3-

digit span 3.40±1.00 4.20±1.70 6.70±5.30 7.70±3.40 0.20 [-3.14, 3.54] 0.05 [-0.83, 0.93]

Backwards 3-digit span

3.40±1.00 7.50±13.00

6.70±5.30 10.40±2.30 -0.40 [-3.74, 2.94] -0.10 [-0.98, 0.78]

Step regularity Gillian, 2009 Backwards

counting from 50 by 1s

276.00±35.00

258.00±38.00

227.00±82.00

139.00±81.00

-70.00 [-120.17, -19.83]

1.28 [0.22, 2.33]*

-89-

Stance time CV

Hsu, 2014 Backwards counting from 100 by 1s

3.13±1.07 19.47±27.4

3.31±0.84 43.31±37.32 23.66 [23.15, 24.17] 23.20 [19.24, 27.16]

Swing time CV Hsu, 2014 Backwards

counting from 100 by 1s

2.47±0.64 13.23±8.77

2.67±0.69 17.24±12.99 6.81 [6.48, 7.14] 10.29 [8.47, 12.10]

Stance period CV Hsu, 2014 Backwards

counting from 100 by 1s

1.8±0.47 5.97±4.08 4.78±5.48 12.3±10.18 3.35 [1.83, 4.87] 1.11 [0.57, 1.66]

Swing period CV Hsu, 2014 Backwards

counting from 100 by 1s

2.17±0.58 6.88±5.91 5.74±7.8 18.95±15.09 8.50 [6.35, 10.65] 1.99 [1.38, 2.60]

Step time variance Konig, 2017 Backwards

counting from 305 by 1s

4.50±4.90 3.90±5.40 6.70±7.10 10.20±9.90 4.10 [0.52, 7.68] 0.66 [0.06, 1.26]

Stride time DFA Lamoth, 2011 Word naming

with 'R' or 'G' 0.87±0.15 0.74±0.15 0.84±0.16 0.84±0.11 0.13 [0.01, 0.25] 0.81 [0.01, 1.62]

SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; DFA=Detrended fluctuation analysis; *=ES was

reversed to indicate change in the direction of other outcome measures. ESs were calculated as standardized mean difference and 95% CI.

-90-

TABLE 2.7 Gait outcome: within study MCI vs. dementia

Author, Year [reference] Task

Control group Cognitive impairment

group Mean Difference (95% CI)

Between group ES (95% CI) Single-task

Mean±SD Dual-task Mean±SD

Single-task Mean±SD

Dual-task Mean±SD

Stride time CV

Muir, 2012

Backwards counting from 100 by 1s

2.59±1.47 4.06±2.37 2.67±1.08 4.86±2.74 0.72 [0.00, 1.44] 0.54 [-0.02, 1.10]

Animal naming 2.59±1.47 8.02±8.88 2.67±1.08 9.04±8.94 0.94 [0.22, 1.66] 0.71 [0.14, 1.27]

Backwards counting from 100 by 7s

2.59±1.47 10.07±9.29 2.67±1.08 12.49±12.33 2.34 [1.62, 3.06] 1.76 [1.11, 2.41]

Step regularity

Gillian, 2009

Backwards counting from 50 by 1s

287.00±29.00 224.00±47.00 227.00±82.00 139.00±81.00 -25.00 [-72.58, 22.58] 0.48 [-0.49, 1.45]*

Step time variance

Konig, 2017

Backwards counting from 305 by 1s

5.70±4.50 6.80±5.30 6.70±7.10 10.20±9.90 2.40 [-0.98, 5.78] 0.40 [-0.18, 0.98]

-91-

MCI=Mild Cognitive Impairment; SD=Standard deviation; ES=Effect size; CI=Confidence interval; CV=Coefficient of variation; *=ES was

reversed to indicate change in the direction of other outcome measures. ESs were calculated as standardized mean difference and 95% CI.

-92-

TABLE 2.8 Stratification of outcomes for meta-analysis

Grouping Stratification Variable Tau2 Q df I2 Z ES Single-task vs. Dual-task

Total

1.32 553.30 84 (P<0.00001) 85% N/A SMD

Cognitive status MCI only 2.21 233.29 29 (P<0.00001) 88% N/A SMD

Dementia only 0.84 269.47 54 (P<0.00001) 80% N/A SMD

Cognitive status and gait outcome

MCI and stride time CV only 5.24 96.03 16 (P<0.00001) 83% N/A MD

Dementia and stride time CV only

6.61 112.18 18 (P<0.00001) 84% N/A MD

Cognitive status, gait outcome and dual-task paradigm

MCI, stride time CV and backwards counting by 1s

7.79 90.76 4 (P<0.00001) 96% N/A MD

MCI, stride time CV, categorical verbal fluency

3.39 74.21 4 (P<0.00001) 95% N/A MD

Dementia, stride time CV, Backwards counting by 1s

169.27 4964.94 6 (P<0.00001) 100% N/A MD

Dementia, stride time CV, categorical verbal fluency

17.43 20.38 2 (P<0.0001) 90% N/A MD

Control vs. Cognitively impaired

Total

3.35 20054.93 40 (P<0.00001) 100% N/A SMD

Cognitive status MCI only 0.21 1174.57 18 (P<0.00001) 98% N/A SMD

Dementia only 58.08 14707.80 21 (P<0.00001) 100% N/A SMD

Cognitive status and gait outcome

MCI and stride time CV only 4.46 245.24 8 (P<0.00001) 97% N/A MD

-93-

Dementia and stride time CV only

42.5 3917.86 12 (P<0.00001) 100% N/A MD

Cognitive status, gait outcome and dual-task paradigm

MCI, stride time CV and backwards counting by 1s

11.21 96.88 2 (P<0.00001) 98% N/A MD

MCI, stride time CV, categorical verbal fluency

1.57 23.3 2 (P<0.00001) 91% N/A MD

Dementia, stride time CV, Backwards counting by 1s

78.01 2845.79 4 (P<0.00001) 100% N/A MD

Dementia, stride time CV, categorical verbal fluency

23.33 75.82 2 (P<0.00001) 97% N/A MD

Within study: MCI vs. Dementia

Total 0 3.74 4 (0.44) 0% 3.06 (0.002)

SMD

MCI=Mild Cognitive Impairment; I2=Measures heterogeneity; Z=Test for overall effect; ES=Effect size; SMD=Standardized mean difference;

MD=Mean difference; CV=Coefficient of variance; P=p value, denoting significance.

-94-

CHAPTER 3: STUDY

The Effects of Dual-Tasking on Gait Dynamics and Cognitive Performance in Adults

with Mild Cognitive Impairment: Contributing Factors and Nonlinear Outcomes

Authors and contributions:

Tess C Hawkins MExPhys, Rebecca Samuel BAppSc, Yorgi Mavros PhD, Nicola Gates PhD,

Guy C Wilson MSc, Nidhi Jain MPH, Jacinda Meiklejohn BS, Henry Brodaty DSc, Wei Wen

PhD, Nalin Singh MBBS, Bernhard T Baune PhD, Chao Suo PhD, Michael K Baker PhD,

Nasim Foroughi PhD, Yi Wang PhD, Perminder S Sachdev PhD, Michael J Valenzuela PhD,

Jeffrey M Hausdorff PhD, and Maria A Fiatarone Singh MD.

Study concept and design: TCH, RS, YM, MAFS and JMH. Acquisition of data: NG, GCW,

NJ, JM, CS, MKB, NF and YW. Analysis and interpretation of data: TCH, RS, YM, MAFS

and JMH. Drafting of the manuscript: TCH, RS, YM, JMH and MAFS. Critical revision of

the manuscript for important intellectual content: TCH, RS, NG, GCW, NJ, JM, WW, MKB,

NF, YW, HB, NS, BTB, CS, PSS, MV, YM, MAFS and JMH. Statistical analysis: TCH, RS

and YM. Obtained funding: HB, WW, NS, BTB, PSS, MV, and MAFS. Administrative,

technical, and material support: NJ, JM, CS. Study supervision: NG, HB, PSS, MV and

MAFS.

-95-

Faculty of Health Sciences

3.1 AUTHOR CONTRIBUTION STATEMENT

Candidate Name: Tess C Hawkins

Degree Title: Master of Applied Science (Research)

Paper Title: The effects of dual-tasking on gait dynamics and cognitive performance in

adults with Mild Cognitive Impairment

As the research supervisor of the above candidate, I confirm that the above candidate has

made the following contributions to the above paper title:

- Conception and design of the research

- Analysis and interpretation of the findings

- Writing the paper and critical appraisal of content

Professor Maria Fiatarone Singh

Discipline of Exercise & Sport Science

Faculty of Health Sciences

The University of Sydney

-96-

3.2 PREAMBLE

The preceding chapter (Chapter 2) detailed the dual-task cost of gait performance in

cognitively impaired older adults. The majority of the studies (92%) reported linear gait

outcome measures only, with nonlinear gait outcome measures reported in just two studies

(8%). The low representation of nonlinear gait dynamics data may distort the understanding

of how impactful dual-tasking is on cognitively impaired older adults. Additionally, the

reporting of characteristics associated dual-task gait dynamics was lacking, which limits the

understanding of what contributes to dual-task cost in cognitively impaired older adults. This

chapter (Chapter 3) presents the findings of a study that used both linear and nonlinear

measures to determine the effects of dual-tasking on gait dynamics in adults with cognitive

impairment, and identified physical, psychosocial, and structural brain characteristics

associated with dual-task cost.

-97-

3.3 ABSTRACT

Objectives

Individuals with Mild Cognitive Impairment (MCI) have increased gait variability. Dual-

tasking can detect interactions between gait dynamics, including variability, and cognition.

We aimed to determine the acute effects of dual-tasking on gait dynamics and cognition in

MCI and to identify associated clinical characteristics.

Design

Acute exposure to dual-tasking during baseline assessment of a randomized controlled trial.

Setting and participants

Ninety-three individuals with MCI (mean age 70±6.8 years; 66.6% female) from the

interventional ‘SMART Study’.

Methods

Cognition, gait, brain Magnetic Resonance Imaging (MRI), muscle strength, aerobic

capacity, body composition, physical and psychosocial function were assessed. Dual-task gait

was measured using force-sensitive insoles to quantify temporal gait dynamics; specifically,

stride time variability and DFA (detrended fluctuation analyses fractal scaling exponent). The

relationship between gait dynamics and cognitive performance was evaluated using linear

mixed models with repeated measures, adjusted for confounders. Linear regression explored

hypothesized mediators of the potential dual-tasking deficits.

Results

Gait dynamics worsened significantly during dual-tasking, with performance decrements in

both stride time variability (p<0.001) and DFA (p=0.001). Lower aerobic capacity and thinner

-98-

posterior cingulate cortex were associated with greater performance decrements in DFA;

whereas smaller hippocampal volume, worse psychological well-being and poorer static

balance were associated with greater performance decrements in stride time variability.

Notably, cognitive performance on the secondary dual-task did not change under dual-task

conditions.

Conclusions/implications

Participants with MCI preserved their cognitive performance at the cost of their gait dynamics

when dual-tasking. We have shown, for the first time that the decrements in dual-tasking gait

are associated with lower aerobic fitness, balance, psychological well-being, and brain

volume in cognitively-relevant areas of the posterior cingulate and hippocampus in MCI; all

of these characteristics are modifiable by exercise. Thus, targeted exercise interventions are

needed to determine the potential plasticity of gait dynamics when stressed in vulnerable

cohorts.

Key words

Dual-task, Gait variability, Gait dynamics, Mild cognitive impairment, and Walking.

-99-

3.4 INTRODUCTION

The prevalence of gait disorders increases with age, affecting up to 35% of community-

dwelling older adults [1]. Increasing gait variability is associated with increased risk of falls

[2] and reduced mobility [3]. Mild Cognitive Impairment (MCI), an intermediate stage

between normal cognition and dementia, is associated with greater gait variability [4] and

double the risk of injurious and multiple falls compared to cognitively normal adults [5].

Dual-tasking is a sensitive method used to investigate interactions between gait variability

and cognitive domains [6, 7], and is associated with increased fall risk amongst older adults

[8]. Dual-tasking impairs gait in individuals with deficits in cognitive function, including

MCI and Alzheimer’s disease [7].

Despite the well-characterized worsening of gait under dual-task conditions, studies

evaluating dual-task associations in older adults with MCI are limited. Identification of

modifiable characteristics associated with gait dynamics under dual-task conditions may lead

to targeted interventions to reduce falls in older adults with MCI. Therefore, we aimed to

determine the effects of dual-tasking on gait dynamics and cognitive performance in adults

with MCI, and to identify physical, psychosocial, and structural brain characteristics

associated with decrements due to dual-tasking (dual-task cost). We hypothesized that there

would be a worsening of both gait and the performance of the secondary cognitive task during

the dual-task condition, and that these reductions would be associated with lower strength,

aerobic capacity, functional performance, psychosocial function and smaller hippocampal

volume and posterior cingulate cortex thickness. These factors were selected a priori due to

their known decrements in cognitive impairment, frailty, or falls [9].

3.5 METHODS

-100-

The complete study protocol for the Study of Mental and Resistance Training (SMART) has

been published [10], with primary [11] and secondary outcomes published [12, 13]. The study

was approved by the Royal Prince Alfred Human Research Ethics Committee (X04-0064),

and written informed consent was obtained from all participants. The study was registered

with the Australia New Zealand Clinical Trials Registry (ACTRN12608000489392).

Participants

One hundred community-dwelling older adults with MCI (Peterson criteria [14]) were

recruited. Two participants could not wear gait monitors due to fused toes, while technical

issues, including incomplete data recording, precluded full gait data in five others. Thus, gait

dynamics data were available for 93 participants. All participants were assessed within the

research clinic space at The University of Sydney Cumberland Campus at Lidcombe in New

South Wales, Australia.

Assessment of cognitive function

Baseline cognitive function has been published [10]. The primary outcome of the SMART

trial was global cognition assessed using the Alzheimer’s Disease Assessment Scale -

Cognition (ADAS-Cog). Attention/speed was assessed via Symbol Digit Modalities Test

(SDMT) and Trail Making Test A. Executive function was assessed by Matrices and

Similarities subtests of the Wechsler Adult Intelligence Scale 3rd Edition (WAIS-III) and

verbal fluency (Controlled Oral Words Association Test (COWAT) and Animal Naming) and

the difference between Trail Making Tests B and A (Trails B – A). Memory tests included

auditory Logical Memory I (immediate) and II (delayed) subtests of the Wechsler Memory

Scale 3rd Edition (WMS-III) and the List Learning subsection of the ADAS-Cog, and visual

via Benton Visual Retention Test-Revised 5th Edition (BVRT-R). Global Domain was the

-101-

average of all z-scores for the gait sample participants (n=93). This included all tests except

List Learning, as it was already included within ADAS-Cog total score.

Assessment of letter fluency at rest and during ambulation

Letter fluency, a subcategory of verbal fluency, was assessed using the COWAT [15] to

measure cognitive performance. Participants were instructed to name as many words

beginning with the letter “F” in 1 minute (FSINGLE), excluding proper nouns, repeated words,

and variations of the same word using a prefix or suffix (e.g., bath and bathing). A score was

calculated by totaling the number of admissible words.

Assessment of gait dynamics

Gait dynamics were assessed similarly to methods previously described [2, 16, 17]. Briefly,

force-sensitive insoles were placed in the participants’ shoes to measure the force applied to

the ground during ambulation. Two sensors were used, one under the heel and another under

the forefoot and toes. A small, lightweight recorder with an on-board A/D converter (12 bit)

was worn on the ankle to sample the output of the insoles at 300Hz and record the data. The

digitized data were transferred to a workstation for analysis, using software that extracts the

initial contact time of each stride [17].

Participants were instructed to walk at their preferred walking speed for two minutes in a

well-lit, indoor hallway with an open path of 25 meters in length. When at the end of the

hallway a large half circle turn was instructed rather than turning on the spot. The first and

last ten seconds of each assessment were removed to minimize starting acceleration or ending

deceleration effects, and a median filter was applied to data points that were three SDs higher

or lower than the median value to remove any outliers due to turns or other irregular gait

-102-

patterns [16]. Participants were instructed to wear habitual, low-heeled shoes, comfortable

clothing, visual and hearing aids, if required.

Gait assessment was performed under two conditions, ‘single-task’ (undistracted walking)

and ‘dual-task’ (distracted walking), in a randomized order with no instruction about task

priority. Stride time variability (coefficient of variation, CV) and detrended fluctuation

analyses fractal scaling exponent (DFA) [18-20] were calculated for both single-task

[(CVSINGLE) and (DFASINGLE)] and dual-task [(CVDUAL) and (DFADUAL)] conditions. To

assess stride-to-stride variability and arrhythmicity of gait, the coefficient of variation (CV)

in each participant’s stride time was calculated using the formula CV = (Standard Deviation

/ Mean) * 100, with lower CV indicating more stable gait. To quantify how the dynamics of

stride times fluctuate and change over time, we applied detrended fluctuation analysis (DFA)

to each participant’s sequence of stride times. Specifically, DFA is a scaling analysis method

used to quantify long-range power-law correlations in signals, and evaluates the fractal

scaling of exponents and the degree of randomness in highly non-stationary physiological

data. DFA eliminates trends in time-series, and can therefore avoid the spurious detection of

correlations from non-stationary artefacts. In general, physiologically healthy systems have

fractal scaling indices between 0.8 and 1.0, with values closer to 0.5 indicating a less healthy

state [21]. The importance of this index is supported by findings among adults with known

altered gait dynamics, in whom a lower fractal scaling index was the only gait dynamics

parameter found to distinguish fallers from non-fallers [21]. The CV and DFA assess and

quantify the changes over time in gait through different methods, the former captures the

magnitude of the changes and the latter captures changes over time. Both are necessary to

assess the totality of gait dynamics and associated factors.

-103-

The secondary task was the COWAT test, with the number of correct “F words” counted in

the first minute of walking (FDUAL). The letter “F” was selected to remain consistent with the

seated COWAT test described above (FSINGLE). No instruction was provided about

prioritizing performance of either the walking or cognitive task but they were encouraged to

do their best during all trials. The words were recorded using a portable recorder (Samsung

YP-U3, Samsung Electronics Co., South Korea) and the recording was subsequently

reviewed by the research assistant who counted the number of admissible words during the

first and second minute. The number of correct words in the first minute of the dual-task gait

condition (FDUAL) was compared to the number of correct words during the FSINGLE condition.

The order was not randomized for the assessment of “F” words, with the dual-task condition

performed 1 week after the seated COWAT. The ‘dual-task cost’ for CV, DFA and F words

were calculated (CVCOST= CVDUAL – CVSINGLE; DFACOST= DFADUAL – DFASINGLE; FCOST =

FDUAL – FSEINGLE).

Assessment of Neuroimaging Outcomes

Details of the neuroimaging assessment outcomes have been published [10, 13]. Magnetic

Resonance Imaging (MRI) data were acquired using a 3.0-Tesla Philips Achieva System

(Achieva, Phillips Healthcare, Best, The Netherlands). Brain structure was assessed using a

T1-weighted whole brain scan (sequence: T1TFE; TR/TE: 6.39/2.9 ms; slice thickness 1.0

mm without gap; field of view: 256 × 256; resolution 1 × 1 mm). Brain volumes were assessed

using 1H-MRS for regional measures: left hippocampus (20 mm M/L, 15 mm D/V, 30 mm

A/P, oriented along the hippocampus) and posterior cingulate grey matter (20 mm M/L, 20

mm D/V, 20 mm A/P) using the PRESS sequence (TE/ TR = 30/2000 ms, 1024 points, 256

averages). Automated and semi-automated computational neuroanatomical analyses were

-104-

performed using a combination of different software packages, in addition to expert manual

tracing of hippocampus.

Assessment of peak strength

Maximal strength testing was assessed via the one-repetition maximum (1RM) on Keiser

pneumatic resistance machines (Keiser Sports Health Equipment, Ltd., Fresno, CA).

Participants’ 1RM was determined on the leg press, knee extension, hip abduction, chest press

and seated row machines. One RM tests were performed twice, one week apart, with the best

performance used.

Assessment of peak aerobic capacity

Aerobic capacity (VO2peak) was determined via indirect calorimetry during a physician-

administered, graded treadmill walking test to volitional fatigue. Methods and data handling

have been previously published [12].

Measures of physical function

Details of the physical function assessments have been published [10]. Static balance was

assessed on one attempt using six different positions (wide stance, narrow stance, semi-

tandem stance, tandem stance, one leg with eyes open and one leg with eyes closed). The

time achieved for each stance was measured, with participants instructed to maintain balance

for 15 seconds. The total static balance score (maximum 90) was calculated by summing the

time for all six positions [22]. All other physical function testing was done in duplicate, with

the better score used in analyses except for habitual gait velocity, for which the average was

used. Habitual and maximal gait velocities were assessed over two metres using an Ultra-

timer (Raymar, Oxfordshire, UK). Participants were instructed to start walking, with the timer

-105-

initiated after the participant had walked 2-meters and the timer automatically stopping as

soon as the participant had walked a further 2-meters. The participant was instructed to walk

to an endpoint 3-meters beyond where the timing would stop. This process was implemented

to avoid the recording of acceleration and deceleration. Dynamic balance was assessed using

the time taken to forward tandem walk over a 3-meter marked course. The instruction to the

participant was to walk as quickly as possible and with as few errors as possible. Errors

included touching the examiner or any object in the assessment environment, stepping

without heel-toe contact and losing balance requiring the support of the examiner. Time was

only stopped when the participant’s whole foot crossed the 3-meter line. Lower extremity

function and power was assessed using the sit-to-stand [23] and stair climb tests [10]. For the

sit-to-stand, participants were instructed to stand up and sit down 5 times as quickly as

possible. Finally, stair climb power (W) was measured by asking participants to ascend a

flight of stairs as quickly as possible. Stair climb power was calculated using the following

formula P (watts) = (M × D) × 9.8/t Where: M = Body mass (kg), D= Vertical distance (m),

D = vertical height of the staircase, and t = Time (s). The 6-minute walk test (6MWT) was

assessed twice, at least one week apart.

Anthropometry and Body Composition

Height, naked body mass and waist circumference were measured as the mean of triplicate

measures after a 12-hour overnight fast. Height was measured to the nearest 0.1cm using

stretch stature with a wall mounted Holtain stadiometer (Holtain Limited, Crymmych Pembs.,

UK). Body mass was measured using a calibrated scale HW-100k & SECA Wedderburn

(>100 kg). Body Mass index (BMI) was calculated by dividing the participants’ body mass

in kilograms by the square of their height in metres. Waist circumference was measured with

Lufkin steel tape measure (W606 PM), using the International Diabetes Federation (IDF)

-106-

protocol [24]. Bioelectrical Impedance Analysis (BIA; RJL Systems, Inc., Clinton, MI, USA)

was used to evaluate body composition. Whole body skeletal muscle mass (kg) [25] and fat

free mass (kg) [26] were calculated using the average resistance and reactance values of three

sequential BIA measures. Further details of the assessments are published [10].

Measures of Psychosocial Function

Details of the assessments used to measure psychosocial function and assess psycho-social

well-being and quality of life have been published [10]. The following tests were used to

measure psychosocial function and assessed psycho-social wellbeing and quality of life via

the Life Satisfaction Scale (LSS) [27], Scale of Psychological Well Being (SPWB) [28],

Quality of Life Scale (QOLS) [29], Physical and Mental Health Short-36 (SF-36), Depression

Anxiety Stress Scale (DASS 21) [30], Memory Awareness Rating Scale – Memory

Functioning Scale (MARS-MF) [31], Duke Social Support Index Scale (DSSIS) [32] and

Life Experience Questionnaire (LEQ).

Statistical Methods

Data were inspected for normality. Normally-distributed data are presented as mean±SD and

non-normally distributed data presented as median (interquartile range). All CV (CVSINGLE,

CVDUAL and CVCOST) variables were log-transformed prior to use in parametric statistics.

Sequential linear regression models (adjusted for age and sex) were constructed including

potential confounders associated with the number of F words during the COWAT test

(FSINGLE) and single- and dual-task gait dynamics (CVSINGLE and DFASINGLE). Next, linear

mixed models with repeated measures were constructed to determine the effect of dual-

tasking on both letter fluency and gait dynamics. The single-task condition (FSINGLE, CVSINGLE

and DFASINGLE) was entered as time point 1, and the ‘dual-task’ condition (FDUAL, CVDUAL

-107-

and DFADUAL) as time point 2. A compound symmetry covariance matrix was used. Models

were adjusted for age and sex, with gait dynamics outcomes further adjusted for the order of

the walking condition. Next, linear regression models were constructed to determine variables

associated with FCOST, CVCOST and DFACOST. Models were adjusted for age, sex, and baseline

score of the dependent variable and order of the condition for the gait dynamics assessment.

All models involving cognitive performance (including FSINGLE and FCOST) were further

adjusted for education, while models involving DFA were further adjusted for the number of

medications, a covariate. Statistical significance was assumed at <0.05 level without

Bonferroni adjustment, as all hypotheses were specified a priori. [33]. All data were analysed

using IBM SPSS (version 24; IBM Corp., Armonk, NY).

3.6 RESULTS

Baseline clinical characteristics of the participants have been published [11]. Data for the

available 93 participants (66.6% women), were similar to the overall cohort (p>0.05 for all

variables). The average age was 70.0±6.8 years, MMSE score 27.5±1.4, and habitual gait

speed 1.21±0.24 m/s, with 17% of participants having a gait speed below 1.0 m/s. Participant

characteristics data are presented in Table 3.1.

Factors associated with gait dynamics

The number of medications prescribed was inversely associated with DFASINGLE (r=-0.23,

p=0.029), but not CVSINGLE (r=-0.03, p=0.750). There were no associations between age, sex,

years of education, smoking status, drinking status or the number of chronic diseases and

either index of gait dynamics. Consequently, all analyses with DFACOST as a dependent

variable were adjusted for the number of medications.

-108-

Dual-Task Cost

Data are presented in Figure 3.1. As hypothesized, gait dynamics worsened significantly

during dual-tasking, with decrements in performance observed for stride time variability

(single-task 2.012 (0.767), dual-task 2.555 (2.227)) and DFA (single-task 0.804±0.151, dual-

task 0.745±0.160). However, contrary to our hypothesis, cognitive performance on the

COWAT did not significantly change under dual-task conditions (single-task 13±5, dual-task

12±4 (Figure 3.1). Changes in letter fluency (FCOST) were not associated with changes in gait

dynamics for either CVCOST (r=-0.14, p=0.171) or DFACOST (r=0.03, p=0.746).

Factors associated with changes in stride time variability during dual-tasking (CVCOST)

Data are presented in Table 3.2. Contrary to our hypotheses, cognitive performance

(executive function, memory, attention and global domains) was not associated with CVCOST

under dual-task conditions (p>0.05). However, higher CVCOST was associated with smaller

(or lower) left hippocampus volume (r=-0.35, p=0.023) with a similar trend for total

hippocampus volume (r=-0.31, p=0.050). Gait dynamics and brain morphology data are

presented in Figure 3.2.

Worse dynamic balance (longer tandem walk time) was directly associated with a higher

CVCOST (r=0.28, p=0.022). However, CVCOST was not associated with body composition,

aerobic capacity, strength or other measures of functional performance.

As hypothesized, psychological well-being was inversely associated with CVCOST, with

higher Environmental Mastery, Personal Growth, Personal Relations, Purpose in Life, Self-

acceptance and total score on the Psychological Wellbeing Scale associated with preservation

of gait under dual-task conditions (p<0.05). Similarly, CVCOST was inversely associated with

-109-

Duke Social Support Index Scale (DSSIS), indicating that the lower level of, and satisfaction

with, social support was associated with a greater gait impairment during dual-task conditions

(p<0.05).

Factors associated with changes in fractal scaling exponent of gait during dual-tasking

(DFACOST)

Contrary to our hypotheses, cognition and psychosocial function were not associated with

DFACOST under dual-task conditions (p>0.05). Notably, as anticipated, greater left (r=0.23,

p=0.026) and total (r=0.25, p=0.015) posterior cingulate cortex thickness as well as better

performances in the 6MWT (r=0.25, p=0.025) and aerobic capacity (r=0.24, p=0.033) were

related to preservation of gait during dual-tasking. However, DFACOST was unrelated to lower

body strength (r=0.25, p=0.063), whole body strength (r=0.25, p=0.079), static balance time

(r=0.18, p=0.075) or tandem walk score (r=-0.12, p=0.254).

3.7 DISCUSSION

The primary finding from this investigation was that in older adults with MCI, resting

cognitive performance on a letter fluency task was preserved under dual-task conditions,

whereas a significant worsening of gait dynamics, both in the magnitude (CV) and time

course (DFA), was observed. Furthermore, worsening of DFA was found to be associated

with lower posterior cingulate cortex thickness and aerobic and walking capacity, while

worsening of stride time variability was associated with smaller hippocampal volume, static

balance and psychological well-being. In contrast, single-task (seated) cognitive function

predicted neither gait variability nor gait dynamic changes during dual-tasking. Although

previous studies [4, 34] have focused on the decrements in gait under dual-task conditions,

-110-

our study has investigated the effects of dual tasking on cognitive performance in MCI, as

well as the mediating role of brain morphology and physical/psychological function in MCI.

We observed preservation of cognitive performance while sacrificing gait dynamics under

dual-task conditions in adults with MCI, which is in agreement with previous studies in

healthy older adults [35] and adults with Parkinson’s disease [36]. This may in part explain

why worsening of gait has previously been associated with recurrent fallers [37] and future

risk of falls [2, 38], and why people with MCI are at a greater risk of falling than their healthy

counterparts [5].

Potential mediators of dual-tasking deficits

Brain morphology

We have reported for the first time that thicker left and total posterior cingulate cortices were

associated with preservation of DFA during dual-tasking. The posterior cingulate cortex

contributes to bilateral lower limb coordination [39] and motor imagery [40], thus, posterior

cingulate thickness reduction with age or MCI could theoretically reduce walking

coordination and result in a more variable gait, particularly during dual-tasking [41], as we

observed in relation to DFA. Additionally, the posterior cingulate cortex assists in directing

the focus of attention [42], which requires information to be integrated [43]. Reduction in

thickness may impact the ability of the posterior cingulate to function; indeed individuals

with clinical disorders associated with posterior cingulate cortex abnormalities have difficulty

regulating the focus of attention [43]. Notably, we have shown that high intensity strength

training increases posterior cingulate thickness [13] and that this change in posterior cingulate

thickness is directly associated with the cognitive benefits of the strength training [13]. Future

investigations are needed to determine whether gait dynamics also improve after robust

-111-

strength training, and whether such changes are related to changes in brain morphology or

other adaptations such as muscle strength or balance.

We also report for the first time the novel finding that lower left and total hippocampal

volume are associated with greater deficits in stride time variability during dual-tasking.

Individuals with MCI have both a higher stride time variability [44, 45] and a lower

hippocampal volume [44] than cognitively healthy individuals. Interestingly, higher stride

time variability has been significantly associated with a lower hippocampal volume in

cognitively healthy individuals during uninterrupted walking [44], however, unexpectedly

not in individuals with MCI [44]. The hippocampus contributes to rhythmicity of locomotion

[46] and is known to atrophy in individuals with MCI faster than in healthy adults [47], which

supports our observed association between lower hippocampal volume and worse gait

dynamics during dual-tasking. It has been postulated that greater hippocampal volume may

compensate for impaired gait dynamics [39] and the hippocampal atrophy of MCI may

mediate diminished ability to regulate gait dynamics under the stressful condition of dual-

tasking in this cohort.

Thus, our novel results for both posterior cingulate cortex and hippocampus provide strong

new evidence of associations between brain morphology, cognitive impairment and gait

variability [48]. We have previously shown that resistance training can significantly increase

posterior cingulate cortex thickness in MCI, and that this increase is associated with the

cognitive benefits of the exercise [13]. This suggests potent avenues to investigate

mechanisms by which exercise may improve gait dynamics in vulnerable cohorts with

cognitive or other neurological impairment.

-112-

Physical fitness

We have shown that higher aerobic capacity and 6-min walking distance are associated with

better preservation of DFA during dual-tasking. High intensity treadmill training has been

shown to improve aerobic capacity and cognition in adults with amnestic MCI [49], whereas

most other low-moderate intensity aerobic interventions have yielded non-significant

findings in this cohort [50]. Progressive resistance training can also improve aerobic capacity

and cognitive function, (mediated by muscle strength gains), as we have previously shown

within this cohort [12]. Notably, given the cross-sectional nature of our analyses, reverse

causality cannot be excluded. Individuals with a less variable gait may be more likely to walk

more frequently and thus have a greater aerobic capacity and functional performance

compared to individuals with a more variable gait pattern. Longitudinal exercise studies are

required to investigate whether improved aerobic/walking capacity will also improve gait

variability during single- and dual-task conditions.

Muscle strength was not related to the preservation of either DFA or stride time variability

during dual-task walking. This differs from data we previously reported showing that higher

muscle strength was related to lower variability measures in community-dwelling older adults

[2] and increased strength predicted less variability in older adults with mild functional

impairment after a multi-modal exercise intervention [51]. Alternatively, others have reported

that stride time variability in older adults with higher level gait disorders was not associated

with muscle strength [21], nor was step time variability improved by a muscle strengthening

intervention [52]. The above studies only measured gait variability under single-task

conditions, hence making this study the first to investigate the relationship between strength

and gait variability under dual-task conditions. Herman and colleagues [21] speculated that

the lack of association between strength and gait variability in some prior studies was due to

-113-

the origin of gait variability being pathologic rather than motor-based, suggesting that frontal

lobe dysfunction not modifiable by strength training may be a cause of gait variability rather

than muscle weakness, which may support our findings.

Psychological well-being

The dual-task cost of stride time variability was greater in those with lower psychological

well-being (SPWB) across the domains of Environmental Mastery, Personal Growth,

Personal Relations, Purpose in Life and Self-acceptance. These data are in agreement with

previous evidence suggesting that worse stride time variability is associated with an increased

fear of falling [21] and fear of falling has been previously associated with physical and mental

limitations as well as social functioning [53]. Consistent with our findings on psychological

well-being, depression is a well-known risk factor for falls and hip fractures [54], independent

from anti-depressant medications, which pose additional fall risk [55]. Whether gait dynamics

and brain morphology decrements in depression underlie this risk is an important area for

future investigation, given the known relationship between hippocampal atrophy and

depression [56] and that our cohort were free from major depression. Additionally, MCI has

been associated with reduced psychological wellbeing [57] and increased falls [38],

supporting a potential link between worsened gait dynamics and poor psychological

wellbeing, but again reverse causality or bi-directional relationships cannot be ruled out.

Limitations

As noted above, reverse causality could explain some of the study outcomes due to the cross-

sectional study design. Also, as noted, the order of the COWAT assessment and dual-task

walking was not randomized. The presentation of the “F” word task at rest prior to dual-

tasking may have produced a learning effect, minimizing the observation of cognitive deficits

-114-

during dual-tasking. Future studies should randomize the sequencing of single- and dual-task

cognitive performance. Finally, appropriately designed studies are warranted to determine

which clinical characteristics related to the dual-task cost remain as independent predictors

following multiple or stepwise regression.

3.8 CONCLUSIONS AND IMPLICATIONS

Older adults with MCI preserved their cognitive performance at the cost of the variability and

dynamics of their gait under dual-task conditions. Novel associations were observed between

worsening of the fractal scaling exponent of gait and posterior cingulate cortical thickness,

while worsening of stride time variability was associated with lower hippocampal volume.

Better aerobic and walking capacity, psychological wellbeing and static balance were also

associated with preservation of gait during a cognitive stressor. Notably, all these factors have

previously been shown to be modifiable with robust exercise modalities in clinical trials.

Thus, longitudinal research is required to determine the extent to which gait dynamics are

also modifiable, and the optimal exercise prescriptions needed to promote optimization of

gait patterns, and ultimately reduce fall risk in vulnerable cohorts.

-115-

3.9 REFERENCES

1. Verghese, J., et al., Epidemiology of Gait Disorders in Community-Residing Older

Adults. Journal of the American Geriatrics Society, 2006. 54(2): p. 255-261.

2. Hausdorff, J.M., D.A. Rios, and H.K. Edelberg, Gait variability and fall risk in

community-living older adults: A 1-year prospective study. Archives of Physical

Medicine and Rehabilitation, 2001. 82(8): p. 1050-1056.

3. Brach, J.S., et al., Gait variability and the risk of incident mobility disability in

community-dwelling older adults. The Journals of Gerontology Series A: Biological

Sciences and Medical Sciences, 2007. 62(9): p. 983-988.

4. Montero-Odasso, M., S.W. Muir, and M. Speechley, Dual-task complexity affects gait

in people with mild cognitive impairment: the interplay between gait variability, dual

tasking, and risk of falls. Archives of Physical Medicine & Rehabilitation, 2012.

93(2): p. 293-299.

5. Delbaere, K., et al., Mild Cognitive Impairment as a Predictor of Falls in Community-

Dwelling Older People. The American Journal of Geriatric Psychiatry, 2012. 20(10):

p. 845-853.

6. Bahureksa, L., et al., The Impact of Mild Cognitive Impairment on Gait and Balance:

A Systematic Review and Meta-Analysis of Studies Using Instrumented Assessment.

Gerontology, 2017. 63(1): p. 67-83.

7. Bragatto, V.S.R., et al., Dual-task during gait between elderly with mild cognitive

impairment and Alzheimer: systematic review. Fisioterapia em Movimento, 2017. 30:

p. 849-857.

8. Beauchet, O., et al., Stops walking when talking: A predictor of falls in older adults?

European Journal of Neurology, 2009. 16(7): p. 786-795.

9. Clegg, A., et al., Frailty in elderly people. The lancet, 2013. 381(9868): p. 752-762.

10. Gates, N.J., et al., Study of Mental Activity and Regular Training (SMART) in at risk

individuals: A randomised double blind, sham controlled, longitudinal trial. BMC

Geriatrics, 2011. 11(1): p. 19.

-116-

11. Fiatarone Singh, M.A., et al., The Study of Mental and Resistance Training (SMART)

study—resistance training and/or cognitive training in mild cognitive impairment: a

randomized, double-blind, double-sham controlled trial. Journal of the American

Medical Directors Association, 2014. 15(12): p. 873-880.

12. Mavros, Y., et al., Mediation of Cognitive Function Improvements by Strength Gains

After Resistance Training in Older Adults with Mild Cognitive Impairment: Outcomes

of the Study of Mental and Resistance Training. Journal of the American Geriatrics

Society, 2017. 65(3): p. 550-559.

13. Suo, C., et al., Therapeutically relevant structural and functional mechanisms

triggered by physical and cognitive exercise. Molecular Psychiatry, 2016: p. doi:

10.1038/mp.2016.19.

14. Petersen, R.C., Mild Cognitive Impairment: Clinical Characterization and Outcome.

The Journal of the American Medical Association, 1999. 281(19): p. 303-308.

15. Kaufman, A.S. and E.O. Lichtenberger, Essentials: of WAIS-III assessment. 1999,

New York: J. Wiley & Sons.

16. Hausdorff, J.M., et al., Increased gait unsteadiness in community-dwelling elderly

fallers. Archives of Physical Medicine and Rehabilitation, 1997. 78(3): p. 278-283.

17. Hausdorff, J.M., Z. Ladin, and J.Y. Wei, Footswitch system for measurement of the

temporal parameters of gait. Journal of Biomechanics, 1995. 28(3): p. 347-351.

18. Hausdorff, J.M., Gait dynamics, fractals and falls: Finding meaning in the stride-to-

stride fluctuations of human walking. Human Movement Science, 2007. 26(4): p. 555-

589.

19. Hausdorff, J.M., et al., Is walking a random walk? Evidence for long-range

correlations in stride interval of human gait. Journal of Applied Physiology, 1995.

78(1): p. 349-358.

20. Hausdorff, J.M., et al., Fractal dynamics of human gait: stability of long-range

correlations in stride interval fluctuations. Journal of Applied Physiology, 1996.

80(5): p. 1448-1457.

-117-

21. Herman, T., et al., Gait instability and fractal dynamics of older adults with a

“cautious” gait: why do certain older adults walk fearfully? Gait & Posture, 2005.

21(2): p. 178-185.

22. Rossiter-Fornoff, J.E., et al., A cross-sectional validation study of the FICSIT common

data base static balance measures. Frailty and Injuries: Cooperative Studies of

Intervention Techniques. The journals of gerontology. Series A, Biological sciences

and medical sciences, 1995. 50(6): p. M291.

23. Guralnik, J.M., et al., A short physical performance battery assessing lower extremity

function: Association with self-reported disability and prediction of mortality and

nursing home admission. Journals of gerontology, 1994. 49(2): p. M85-M94.

24. Alberti, K.G., P.Z. Zimmet, and J.M. Shaw, Metabolic syndrome-- a new world-wide

definition. A Consensus Statement from the International Diabetes Federation.

Diabetic medicine : A journal of the British Diabetic Association, 2006. 23(5): p. 469-

480.

25. Jansen, I., et al., Estimation of skeletal muscle mass by bioelectrical impedance

analysis. Journal of Applied Physiology, 2000. 89(2): p. 465-471.

26. Lukaski, H., et al., Validation of tetrapolar bioelectrical impedance method to assess

human body composition. Journal of Applied Physiology, 1986. 60(4): p. 1327-1332.

27. Andrews, F.M. and S.B. Withey, Social indicators of well-being: Americans'

perceptions of life quality. 1976, New York: Plenum Press.

28. Ryff, C.D. and C.L.M. Keyes, The Structure of Psychological Well-Being Revisited.

Journal of personality and social psychology, 1995. 69(4): p. 719-727.

29. Burckhardt, C.S. and K.L. Anderson, The Quality of Life Scale (QOLS): Reliability,

validity, and utilization. Health and Quality of Life Outcomes, 2003. 1(1): p. 60-60.

30. Ng, F., et al., The validity of the 21-item version of the Depression Anxiety Stress

Scales as a routine clinical outcome measure. Acta Neuropsychiatrica, 2007. 19(5):

p. 304-310.

-118-

31. Clare, L., et al., Assessing awareness in early-stage Alzheimer's disease: Development

and piloting of the Memory Awareness Rating Scale. Neuropsychological

Rehabilitation, 2002. 12(4): p. 341-362.

32. Koenig, H.G., et al., Abbreviating the Duke Social Support Index for use in

chronically ill elderly individuals. Psychosomatics, 1993. 34(1): p. 61-69.

33. Nakagawa, S., A farewell to Bonferroni: the problems of low statistical power and

publication bias. Behavioral Ecology, 2004. 15(6): p. 1044-1045.

34. Nascimbeni, A., et al., Dual task-related gait changes in patients with mild cognitive

impairment. Functional Neurology, 2015. 30(1): p. 59-65.

35. Corp, D.T., et al., Reduced motor cortex inhibition and a ‘cognitive-first’

prioritisation strategy for older adults during dual-tasking. Experimental

Gerontology, 2018. 113: p. 95-105.

36. Bloem, B.R., et al., The "posture second" strategy: A review of wrong priorities in

Parkinson's disease. Journal of the Neurological Sciences, 2006. 248(1-2): p. 196-

204.

37. Brach, J.S., et al., Too much or too little step width variability is associated with a fall

history in older persons who walk at or near normal gait speed. Journal of

NeuroEngineering and Rehabilitation, 2005. 2: p. 21-21.

38. Callisaya, M.L., et al., Gait, gait variability and the risk of multiple incident falls in

older people: a population-based study. Age and Ageing, 2011. 40(4): p. 481-487.

39. Tian, Q., et al., The brain map of gait variability in aging, cognitive impairment and

dementia—a systematic review. Neuroscience & Biobehavioral Reviews, 2017. 74: p.

149-162.

40. Rosano, C., et al., Special ArticleGait Measures Indicate Underlying Focal Gray

Matter Atrophy in the Brain of Older Adults. The Journals of Gerontology: Series A,

2008. 63(12): p. 1380-1388.

41. Doi, T., et al., Gray matter volume and dual-task gait performance in mild cognitive

impairment. Brain Imaging and Behavior, 2017. 11(3): p. 887-898.

-119-

42. Leech, R., et al., Fractionating the Default Mode Network: Distinct Contributions of

the Ventral and Dorsal Posterior Cingulate Cortex to Cognitive Control. The Journal

of Neuroscience, 2011. 31(9): p. 3217-3224.

43. Leech, R. and D.J. Sharp, The role of the posterior cingulate cortex in cognition and

disease. Brain, 2014. 137(1): p. 12-32.

44. Beauchet, O., et al., Hippocampal volume, early cognitive decline and gait variability:

Which association? Experimental Gerontology, 2015. 61: p. 98-104.

45. Muir, S.W., et al., Gait assessment in mild cognitive impairment and Alzheimer's

disease: The effect of dual-task challenges across the cognitive spectrum. Gait and

Posture., 2012.

46. Zimmerman, M.E., et al., MRI- and MRS-Derived Hippocampal Correlates of

Quantitative Locomotor Function in Older Adults. Brain Research, 2009. 1291: p. 73-

81.

47. Leal, S.L. and M.A. Yassa, Perturbations of neural circuitry in aging, mild cognitive

impairment, and Alzheimer's disease. Ageing Research Reviews, 2013. 12(3): p. 823-

831.

48. Hamacher, D., et al., Brain activity during walking: A systematic review.

Neuroscience and Biobehavioral Reviews, 2015. 57: p. 310-327.

49. Baker, L.D., et al., Effects of Aerobic Exercise on Mild Cognitive Impairment: A

Controlled Trial. Archives of Neurology, 2010. 67(1): p. 71-79.

50. Gates, N., et al., The effect of exercise training on cognitive function in older adults

with mild cognitive impairment: a meta-analysis of randomized controlled trials. The

American Journal of Geriatric Psychiatry, 2013. 21(11): p. 1086-1097.

51. Hausdorff, J.M., et al., Etiology and modification of gait instability in older adults: a

randomized controlled trial of exercise. Journal of Applied Physiology, 2001. 90(6):

p. 2117-2129.

52. Schwenk, M., et al., Improvements in gait characteristics after intensive resistance

and functional training in people with dementia: a randomised controlled trial. BMC

Geriatrics, 2014. 14(1): p. 73.

-120-

53. Meulen, E., et al., Effect of Fall‐Related Concerns on Physical, Mental, and Social

Function in Community‐Dwelling Older Adults: A Prospective Cohort Study. Journal

of the American Geriatrics Society, 2014. 62(12): p. 2333-2338.

54. Whooley, M.A., et al., Depression, Falls, and Risk of Fracture in Older Women.

Archives of Internal Medicine, 1999. 159(5): p. 484-490.

55. Kvelde, T., et al., Depressive symptoms increase fall risk in older people, independent

of antidepressant use, and reduced executive and physical functioning. Archives of

Gerontology and Geriatrics, 2015. 60(1): p. 190-195.

56. Sheline, Y.I., et al., Hippocampal atrophy in recurrent major depression. Proceedings

of the National Academy of Sciences, 1996. 93(9): p. 3908-3913.

57. Gates, N., et al., Psychological well-being in individuals with mild cognitive

impairment. Clinical Interventions In Aging, 2014. 9: p. 779-792.

-121-

FIGURE LEGENDS FIGURE 3.1: Graphed results under single-task and dual-task walking conditions for (a)

DFA, (b) Stride time variability and (c) Letter fluency. Data for (a) and (c) are presented as

mean±SD and data for (b) are presented as median (IQR) with all adjusted pairwise. DFA is

detrended fluctuation analysis fractal scaling exponent.

FIGURE 3.2: Graphed results of gait dynamics and MRI associations for (a) CVCOST and left

hippocampal volume, (b) DFACOST and left posterior cingulate thickness, and (c) DFACOST

and total posterior cingulate thickness. CVCOST is the dual-task cost of stride time variability.

DFACOST is the dual-task cost of DFA, detrended fluctuation analysis fractal scaling

exponent. Cost variables are calculated by subtracting the single-task from the dual-task.

-122-

FIGURE 3.1 Gait dynamics and cognitive performance under single-task and dual-task

walking conditions

-123-

Data for (a) and (c) are presented as mean±SD and data for (b) are presented as median (IQR)

with all adjusted pairwise. DFA is detrended fluctuation analysis fractal scaling exponent.

Higher DFA and lower CV represent a less variable gait.

-124-

FIGURE 3.2 Relationship between gait dynamics and brain morphology

-125-

CVCOST is the dual-task cost of stride time variability. DFACOST is the dual-task cost of DFA,

detrended fluctuation analysis fractal scaling exponent. Cost variables are calculated by

subtracting the single-task from the dual-task Greater decrements during dual-tasking are

indicated by more positive CV cost and more negative DFA cost.

-126-

TABLE 3.1 Participant characteristics

Characteristics Value

Demographics

Age (yrs) 70.0 ± 6.8

Sex: Female (%) 66.7

BMI (kg/ m2) 26.87 ± 4.89

Education (yrs) 13 ± 3

Total alcoholic drinks / week 3 (9)

Current smoker (%) 2.2

Ex-smoker (%) 38.7

Health status

Medications/day 4 (4)

Number of chronic diseases (%) 2.9 ± 1.7

Osteoarthritis (%) 71.0

Hypertension (%) 40.9

Diabetes (%) 11.8

Gout (%) 4.3

Depressive episodes in the past 5 years 0 (0)

Habitual gait speed (m/s) 1.22 ± 0.24

Results reported in mean±SD or Median (IQR); SD=Standard deviation; IQR=Interquartile

Range; yrs=years; %=percent; BMI=body mass index; kg=kilogram; m= meter; s=second.

-127-

TABLE 3.2 Factors significantly associated with changes in at least one measure of gait

variability and dynamics during dual-tasking

Associated factors

Demographics

CVCOST DFACOST

Age (yrs) -0.032 -0.201*

Psychosocial Assessments

SPWB Environmental Mastery -0.276* -0.091

SPWB Personal Growth -0.217* -0.156

SPWB Personal Relations -0.258* -0.137

SPWB Purpose in Life -0.248* -0.064

SPWB Self-Acceptance -0.244* -0.002

SPWB Total Score -0.303* -0.114

DSSIS (11-33) -0.243* -0.104

Functional Status and Physical Performance

VO2 peak (mL/kg/min) 0.037 0.242*

Best Tandem Walk (s) 0.275* -0.119

6 Minute Walk Distance (m) 0.050 0.245*

Brain MRI Thicknesses and Volumes

Left Hippocampus Volume (mm3) -0.347* -0.184

Left Posterior Cingulate Cortex Thickness

(mm) 0.240 0.228*

Total Posterior Cingulate Cortex Thickness

(mm) 0.246 0.254*

Results reported as r value; yrs=years; kg=kilograms; m= meters; mL=milliliters;

min=minutes; s=seconds; mm=millimeters mm3=millimeter cubed, SPWB=Scale of

-128-

Psychological Well Being; DSSIS=Duke Social Support Index Scale; *=significant p value

(p<0.05).

-129-

CHAPTER 4: CONCLUSIONS AND IMPLICATIONS

4.1 CONCLUSIONS

This thesis has demonstrated a link between cognitive impairment status and the dual-task

cost of gait dynamics, with a particular focus on modifiable and non-modifiable

characteristics within this population. The systematic review (Chapter 2) indicated that gait

dynamics in cognitively impaired older adults are worse when walking under dual-task

conditions compared to single-task conditions. Additionally, the dual-task cost of gait

dynamics is larger for cognitively impaired older adults than healthy older adults. The acute

dual-task study focused on older adults with MCI (Chapter 3) and showed that cognitive

performance was preserved at the cost of worsening their dual-task gait dynamics.

Furthermore, significant and biologically plausible associations were identified between

dual-task gait performance and aerobic capacity, functional performance, psychosocial

function, and brain morphology.

The systematic review (Chapter 2) was aimed at investigating the effect of dual-task walking

on changes in gait dynamics, with respect to differences between cognitive pathology

diagnosis, dual-task paradigms and gait dynamic outcomes in cognitively impaired older

adults. Additionally, this review aimed to identify potentially modifiable and non-modifiable

characteristics of gait dynamics in cognitively impaired older adults. The review found that

gait dynamics are worsened under dual-task conditions compared to single-task conditions in

older adults with both MCI and dementia. The dual-task costs of gait dynamics are further

increased in cognitively impaired older adults compared to healthy older adults, with greater

decrements in gait dynamics observed with more severe cognitive impairment [i.e.,

dementia/Alzheimer’s disease (AD) vs. Mild Cognitive Impairment (MCI)].

-130-

The review also highlighted major limitations within the current literature. Specifically, the

studies were too varied in their methodology to determine if there is a superior dual-task

protocol for measuring gait or falls risk. Studies commonly reported linear measures of gait

dynamics, however, nonlinear measures were only reported twice. Linear and nonlinear

measures of gait dynamics use different methods to assess gait and different aspects of

changes in the gait cycle over time. Both types of outcomes are necessary to fully understand

gait dynamics and associated factors. Additionally, there are MCI and dementia sub-types,

e.g., amnestic MCI (a-MCI) and non-amnestic MCI (na-MCI), and Alzheimer’s disease with

impaired executive function (IEF), extrapyramidal signs (EPS) or frontotemporal dementia

(FTD). There is limited research within these sub-types, and future research should

investigate differences between sub-types with larger sample sizes and varying levels of task

complexity. Furthermore, falls history and fear of falling were not well documented and thus

definitive conclusions about relationships between falls risk and dual-task gait dynamics were

unable to be drawn.

The acute dual-task study (Chapter 3) was aimed at determining the extent of changes in gait

and the performance of a secondary task during dual-task walking using a linear and nonlinear

measure of gait. As identified by the systematic review, there are only two previous studies

published to our knowledge that used nonlinear measures of gait dynamics to quantify the

cost of a dual-task. Additionally, our study sought to explore associations with dual-task

performance and to identify any characteristics that may be modifiable via preventative

therapeutic strategies. These characteristics were identified prior to the study, based on the

existing literature in healthy older adults, gait performance and dual-task walking [1-5], and

included strength, aerobic capacity, functional performance, psychosocial function, and brain

-131-

morphology. Identifying such characteristics associated with dual-task gait performance may

assist in the development of optimal interventions to reduce gait dynamics under single- and

dual-task conditions. The results showed that larger hippocampal volume, thicker posterior

cingulate cortex, higher aerobic and walking capacity, greater psychological wellbeing and

greater dynamic balance were each associated with preservation of gait during dual-task

conditions. These factors have previously been shown to be modifiable with robust exercise

modalities in clinical trials [1, 6]. More research is required to determine the extent of which

gait dynamics are modifiable to support the creation of strategies to improve gait dynamics

and reduce falls in older adults with MCI and other conditions associated with gait variability

and dynamics.

4.2 IMPLICATIONS AND FUTURE DIRECTIONS

As stated, additional longitudinal research is required to determine the most important and

modifiable characteristics associated with gait dynamics under dual-task conditions, as we

cannot attribute causality to the observations we made in our MCI cohort. Dual-task costs

with regard to gait dynamics have been well characterized, however, which physiological and

neuropsychological attributes contribute independently and most consistently to increased

dual-task cost and the future risk of impaired gait dynamics have not been fully defined. If

modifiable characteristics can be determined, interventions could be developed to target these

specific characteristics to improve gait performance and reduce risk of falls. Current research

is being conducted to determine the link between future falls and worse gait dynamics to

understand what information is needed for clinical assessment in the reduction of falls [7].

Our results suggest that not only physiological, but psychosocial aspects of aging need to be

considered in falls risk assessments. A greater dual-task cost of gait dynamics has been

-132-

associated with lower psychosocial well-being and increased fear of falling [8]. The known

links between fear of falling and physical and mental limitations, reduced social functioning

[9], activity restriction, increased anxiety and depression and reduced physical activity levels

[10] promote the need for clinicians to assess falls risk using a multifactorial assessment tool.

Additionally, associations between depression and prevalence of dementia have been shown

[11], which adds further complexity to the relationship between aging, cognition, gait

dynamics and falling. Thus there is need for clinicians to incorporate a psychosocial

component to the fall risk battery of assessments to determine current well-being, fear of

falling and mood.

In the available literature, there is a lack of consensus regarding the methods used to analyze

gait dynamics [12], which prevents the development of comprehensive recommendations for

dual-task gait testing procedures for use in clinical practice [7]. The development and

utilization of consensus guidelines on gait analysis protocols could identify and stratify fall

risk in cognitively impaired older adults as well as other cohorts at risk [7, 13]. Additionally,

standardizing research methodologies would improve the overall understanding of dual-task

gait performance changes [14] and assist in the identification of risk profiled for clinically

relevant gait dynamics.

A recent review has recommended implementation of combined interventions in order to

improve overall gait for older adults with MCI or early dementia [15]. The term ‘combined

intervention’ refers to the inclusion of strength, balance, and functional mobility training,

along with executive function training [15]. This is concordant with the findings of our acute

dual-task study (Chapter 3) which highlighted specific characteristics that were associated

with dual-task gait performance and have also shown to be modifiable with robust exercise

-133-

modalities in clinical trials. To adequately determine the modifiable and non-modifiable

characteristics of dual-task gait dynamics, more well-designed longitudinal studies and

controlled trials with adequately powered samples are needed, and the independent predictive

value of the clinical characteristics we identified should be explored in multivariable

regression or structural equation modelling. Additionally, more longitudinal research is

required to determine the ideal prescriptions (including exercise and/or other interventions)

needed to promote optimization of gait patterns under single- and dual-task conditions to

enhance and preserve mobility and functional independence in older adults, while minimizing

the risk of falls and related adverse events.

-134-

4.3 REFERENCES

1. Baker, L.D., et al., Effects of Aerobic Exercise on Mild Cognitive Impairment: A

Controlled Trial. Archives of Neurology, 2010. 67(1): p. 71-79.

2. Beauchet, O., et al., Relationship between dual-task related gait changes and intrinsic

risk factors for falls among transitional frail older adults. Aging Clinical and

Experimental Research, 2005. 17(4): p. 270-275.

3. Doi, T., et al., Gray matter volume and dual-task gait performance in mild cognitive

impairment. Brain Imaging and Behavior, 2017. 11(3): p. 887-898.

4. Hausdorff, J.M., et al., Etiology and modification of gait instability in older adults: a

randomized controlled trial of exercise. Journal of Applied Physiology, 2001. 90(6):

p. 2117-2129.

5. Hausdorff, J.M., et al., Dual-Task Decrements in Gait: Contributing Factors Among

Healthy Older Adults. Journals of Gerontology Series a-Biological Sciences and

Medical Sciences, 2008. 63(12): p. 1335-1343.

6. Mavros, Y., et al., Mediation of Cognitive Function Improvements by Strength Gains

After Resistance Training in Older Adults with Mild Cognitive Impairment: Outcomes

of the Study of Mental and Resistance Training. Journal of the American Geriatrics

Society, 2017. 65(3): p. 550-559.

7. Muir-Hunter, S.W. and J.E. Wittwer, Dual-task testing to predict falls in community-

dwelling older adults: a systematic review. Physiotherapy, 2016. 102(1): p. 29-40.

8. Herman, T., et al., Gait instability and fractal dynamics of older adults with a

“cautious” gait: why do certain older adults walk fearfully? Gait & Posture, 2005.

21(2): p. 178-185.

9. Meulen, E., et al., Effect of Fall‐Related Concerns on Physical, Mental, and Social

Function in Community‐Dwelling Older Adults: A Prospective Cohort Study. Journal

of the American Geriatrics Society, 2014. 62(12): p. 2333-2338.

10. Painter, J.A., et al., Fear of Falling and Its Relationship With Anxiety, Depression,

and Activity Engagement Among Community-Dwelling Older Adults. American

Journal of Occupational Therapy, 2012. 66(2): p. 169-176.

-135-

11. Bennett, S. and A.J. Thomas, Depression and dementia: Cause, consequence or

coincidence? Maturitas, 2014. 79(2): p. 184-190.

12. Lord, S., et al., Gait variability in older adults: A structured review of testing protocol

and clinimetric properties. Gait & Posture, 2011. 34(4): p. 443-450.

13. Modarresi, S., et al., Gait parameters and characteristics associated with increased

risk of falls in people with dementia: a systematic review. International

Psychogeriatrics, 2018: p. 1-17.

14. Al-Yahya, E., et al., Cognitive motor interference while walking: A systematic review

and meta-analysis. Neuroscience and Biobehavioral Reviews, 2011. 35(3): p. 715-

728.

15. Zhang, W., et al., Interventions to Improve Gait in Older Adults with Cognitive

Impairment: A Systematic Review. Journal of the American Geriatrics Society, 2019.

67(2): p. 381-391.

-136-

BIBLIOGRAPHY

Al-Yahya, E., Dawes, H., Smith, L., Dennis, A., Howells, K., & Cockburn, J. (2011).

Cognitive motor interference while walking: A systematic review and meta-analysis.

Neuroscience and Biobehavioral Reviews, 35(3), 715-728.

Alberti, K. G., Zimmet, P. Z., & Shaw, J. M. (2006). Metabolic syndrome-- a new world-

wide definition. A Consensus Statement from the International Diabetes Federation.

Diabetic medicine : A journal of the British Diabetic Association, 23(5), 469-480.

Allali, G., Annweiler, C., Blumen, H. M., Callisaya, M. L., De Cock, A. M., Kressig, R. W.,

. . . Beauchet, O. (2016). Gait phenotype from mild cognitive impairment to moderate

dementia: results from the GOOD initiative. European Journal of Neurology, 23(3), 527-

541.

Allali, G., Assal, F., Kressig, R. W., Dubost, V., Herrmann, F. R., & Beauchet, O. (2008).

Impact of impaired executive function on gait stability. Dementia & Geriatric Cognitive

Disorders, 26(4), 364-369.

Allali, G., Dubois, B., Assal, F., Lallart, E., de Souza, L. C., Bertoux, M., . . . Beauchet, O.

(2010). Frontotemporal dementia: pathology of gait? Movement Disorders, 25(6), 731-

737.

Allali, G., Kressig, R. W., Assal, F., Herrmann, F. R., Dubost, V., & Beauchet, O. (2007).

Changes in gait while backward counting in demented older adults with frontal lobe

dysfunction. Gait & Posture, 26(4), 572-576.

Allali, G., Launay, C. P., Blumen, H. M., Callisaya, M. L., De Cock, A.-M., Kressig, R. W.,

. . . Biomathics, C. (2017). Falls, Cognitive Impairment, and Gait Performance: Results

From the GOOD Initiative. Journal of the American Medical Directors Association,

18(4), 335-340.

-137-

Andrews, F. M., & Withey, S. B. (1976). Social indicators of well-being: Americans'

perceptions of life quality. New York: Plenum Press.

Arif, M., Ohtaki, Y., Nagatomi, R., & Inooka, H. (2004). Estimation of the effect of cadence

on gait stability in young and elderly people using approximate entropy technique.

Measurement Science Review, 4(2), 29-40.

Auvinet, B., Touzard, C., Montestruc, F., Delafond, A., & Goeb, V. (2017). Gait disorders in

the elderly and dual task gait analysis: a new approach for identifying motor phenotypes.

Journal of Neuroengineering & Rehabilitation, 14(1), 7.

Ayers, E. I., Tow, A. C., Holtzer, R., & Verghese, J. (2014). Walking while Talking and Falls

in Aging. Gerontology, 60(2), 108-113.

Bahureksa, L., Najafi, B., Saleh, A., Sabbagh, M., Coon, D., Mohler, M. J., & Schwenk, M.

(2017). The Impact of Mild Cognitive Impairment on Gait and Balance: A Systematic

Review and Meta-Analysis of Studies Using Instrumented Assessment. Gerontology,

63(1), 67-83.

Baker, L. D., Frank, L. L., Foster-Schubert, K., Green, P. S., Wilkinson, C. W., McTiernan,

A., . . . Craft, S. (2010). Effects of Aerobic Exercise on Mild Cognitive Impairment: A

Controlled Trial. Archives of Neurology, 67(1), 71-79.

Beauchet, O., Annweiler, C., Callisaya, M. L., De Cock, A.-M., Helbostad, J. L., Kressig, R.

W., . . . Allali, G. (2016). Poor Gait Performance and Prediction of Dementia: Results

From a Meta-Analysis. Journal of the American Medical Directors Association, 17(6),

482-490.

Beauchet, O., Annweiler, C., Dubost, V., Allali, G., Kressig, R. W., Bridenbaugh, S., . . .

Herrmann, F. R. (2009). Stops walking when talking: A predictor of falls in older adults?

European Journal of Neurology, 16(7), 786-795.

-138-

Beauchet, O., Dubost, V., Herrmann, F., Rabilloud, M., Gonthier, R., & Kressig, R. W.

(2005). Relationship between dual-task related gait changes and intrinsic risk factors for

falls among transitional frail older adults. Aging Clinical and Experimental Research,

17(4), 270-275.

Beauchet, O., Freiberger, E., Annweiler, C., Kressig, R., Herrmann, F., & Allali, G. (2011).

Test-retest reliability of stride time variability while dual tasking in healthy and

demented adults with frontotemporal degeneration. Journal of NeuroEngineering and

Rehabilitation, 8(1).

Beauchet, O., Launay, C., Allali, G., Herrmann, F., & Annweiler, C. (2014). Gait Changes

with Anti-Dementia Drugs: A Prospective, Open-Label Study Combining Single and

Dual Task Assessments in Patients with Alzheimer's Disease. Drugs & Aging, 31(5),

363-372.

Beauchet, O., Launay, C. P., Annweiler, C., & Allali, G. (2015). Hippocampal volume, early

cognitive decline and gait variability: Which association? Experimental Gerontology, 61,

98-104.

Beauchet, O., Launay, C. P., Sekhon, H., Barthelemy, J.-C., Roche, F., Chabot, J., . . . Allali,

G. (2017). Association of increased gait variability while dual tasking and cognitive

decline: results from a prospective longitudinal cohort pilot study. GeroScience, 39(4),

439-445.

Bennett, S., & Thomas, A. J. (2014). Depression and dementia: Cause, consequence or

coincidence? Maturitas, 79(2), 184-190.

Bjelland, I., Dahl, A. A., Haug, T. T., & Neckelmann, D. (2002). The validity of the Hospital

Anxiety and Depression Scale: an updated literature review. Journal of psychosomatic

research, 52(2), 69-77.

-139-

Blake, A., Morgan, K., Bendall, M., Dallosso, H., Ebrahim, S., Arie, T., . . . Bassey, E. (1988).

Falls by elderly people at home: prevalence and associated factors. Age and ageing,

17(6), 365-372.

Bloem, B. R., Grimbergen, Y. A. M., van Dijk, J. G., & Munneke, M. (2006). The "posture

second" strategy: A review of wrong priorities in Parkinson's disease. Journal of the

Neurological Sciences, 248(1-2), 196-204.

Booth, V., Hood, V., & Kearney, F. (2016). Interventions incorporating physical and

cognitive elements to reduce falls risk in cognitively impaired older adults: A systematic

review. JBI Database of Systematic Reviews and Implementation Reports, 14(5), 110-

135.

Brach, J. S., Berlin, J. E., VanSwearingen, J. M., Newman, A. B., & Studenski, S. A. (2005).

Too much or too little step width variability is associated with a fall history in older

persons who walk at or near normal gait speed. Journal of NeuroEngineering and

Rehabilitation, 2, 21-21.

Brach, J. S., Studenski, S. A., Perera, S., VanSwearingen, J. M., & Newman, A. B. (2007).

Gait variability and the risk of incident mobility disability in community-dwelling older

adults. The Journals of Gerontology Series A: Biological Sciences and Medical Sciences,

62(9), 983-988.

Bradley, C., & Harrison, J. (2007). Hospitalisations due to falls in older people, Australia,

2003-2004 Injury research and statistics series no. 32 (Vol. Cat. no. INJCAT 96).

Canberra: AIHW.

Burckhardt, C. S., & Anderson, K. L. (2003). The Quality of Life Scale (QOLS): Reliability,

validity, and utilization. Health and Quality of Life Outcomes, 1(1), 60-60.

-140-

Byun, S., Han, J. W., Kim, T. H., Kim, K., Kim, T. H., Park, J. Y., . . . Kim, K. W. (2018).

Gait Variability Can Predict the Risk of Cognitive Decline in Cognitively Normal Older

People. Dementia and Geriatric Cognitive Disorders, 251-261.

Callisaya, M. L., Blizzard, L., Schmidt, M. D., Martin, K. L., McGinley, J. L., Sanders, L.

M., & Srikanth, V. K. (2011). Gait, gait variability and the risk of multiple incident falls

in older people: a population-based study. Age and ageing, 40(4), 481-487.

Camicioli, R., Bouchard, T., & Licis, L. (2006). Dual-tasks and walking fast: relationship to

extra-pyramidal signs in advanced Alzheimer disease. Journal of the Neurological

Sciences, 248(1-2), 205-209.

Camicioli, R., & Licis, L. (2004). Motor impairment predicts falls in specialized Alzheimer

care units. Alzheimer Disease & Associated Disorders, 18(4), 214-218.

Clare, L., Wilson, B. A., Carter, G., Roth, I., & Hodges, J. R. (2002). Assessing awareness in

early-stage Alzheimer's disease: Development and piloting of the Memory Awareness

Rating Scale. Neuropsychological Rehabilitation, 12(4), 341-362.

Clegg, A., Young, J., Iliffe, S., Rikkert, M. O., & Rockwood, K. (2013). Frailty in elderly

people. The Lancet, 381(9868), 752-762.

Cohen, J. (1988). Statistical power analysis for the behavioural sciences (Second edition ed.).

Hillside, NJ, USA: Lawrence Erlbaum Associates.

Corp, D. T., Youssef, G. J., Clark, R. A., Gomes-Osman, J., Yücel, M. A., Oldham, S. J., . .

. Rogers, M. A. (2018). Reduced motor cortex inhibition and a ‘cognitive-first’

prioritisation strategy for older adults during dual-tasking. Experimental Gerontology,

113, 95-105.

D'Esposito, M., Postle, B. R., Ballard, D., & Lease, J. (1999). Maintenance versus

Manipulation of Information Held in Working Memory: An Event-Related fMRI Study.

Brain and Cognition, 41(1), 66-86.

-141-

Delbaere, K., Kochan, N. A., Close, J. C. T., Menant, J. C., Sturnieks, D. L., Brodaty, H., . .

. Lord, S. R. (2012). Mild Cognitive Impairment as a Predictor of Falls in Community-

Dwelling Older People. The American Journal of Geriatric Psychiatry, 20(10), 845-853.

Doi, T., Blumen, H. M., Verghese, J., Shimada, H., Makizako, H., Tsutsumimoto, K., . . .

Suzuki, T. (2017). Gray matter volume and dual-task gait performance in mild cognitive

impairment. Brain Imaging and Behavior, 11(3), 887-898.

Downs, S. H., & Black, N. (1998). The feasibility of creating a checklist for the assessment

of the methodological quality both of randomised and non-randomised studies of health

care interventions. Journal of Epidemiology and Community Health, 52(6), 377-384.

Fiatarone Singh, M. A., Gates, N., Saigal, N., Wilson, G. C., Meiklejohn, J., Brodaty, H., . .

. Valenzuela, M. (2014). The Study of Mental and Resistance Training (SMART)

study—resistance training and/or cognitive training in mild cognitive impairment: a

randomized, double-blind, double-sham controlled trial. Journal of the American

Medical Directors Association, 15(12), 873-880.

Florence, C. S., Bergen, G., Atherly, A., Burns, E., Stevens, J., & Drake, C. (2018). Medical

Costs of Fatal and Nonfatal Falls in Older Adults. Journal of the American Geriatrics

Society, 66(4), 693-698.

Gates, N., Singh, M. A. F., Sachdev, P. S., & Valenzuela, M. (2013). The effect of exercise

training on cognitive function in older adults with mild cognitive impairment: a meta-

analysis of randomized controlled trials. The American Journal of Geriatric Psychiatry,

21(11), 1086-1097.

Gates, N., Valenzuela, M., Sachdev, P. S., & Fiatarone Singh, M. A. (2014). Psychological

well-being in individuals with mild cognitive impairment. Clinical Interventions in

Aging, 9, 779-792.

-142-

Gates, N. J., Valenzuela, M., Sachdev, P. S., Singh, N. A., Baune, B. T., Brodaty, H., . . .

Fiatarone Singh, M. A. (2011). Study of Mental Activity and Regular Training (SMART)

in at risk individuals: A randomised double blind, sham controlled, longitudinal trial.

BMC Geriatrics, 11(1), 19.

Gillain, S., Drame, M., Lekeu, F., Wojtasik, V., Ricour, C., Croisier, J. L., . . . Petermans, J.

(2016). Gait speed or gait variability, which one to use as a marker of risk to develop

Alzheimer disease? A pilot study. Aging-Clinical & Experimental Research, 28(2), 249-

255.

Gillain, S., Warzee, E., Lekeu, F., Wojtasik, V., Maquet, D., Croisier, J. L., . . . Petermans, J.

(2009). The value of instrumental gait analysis in elderly healthy, MCI or Alzheimer's

disease subjects and a comparison with other clinical tests used in single and dual-task

conditions. Annals of Physical & Rehabilitation Medicine, 52(6), 453-474.

Gschwind, Y., Bridenbaugh, S., Reinhard, S., Granacher, U., Monsch, A., & Kressig, R.

(2017). Ginkgo biloba special extract LI 1370 improves dual-task walking in patients

with MCI: a randomised, double-blind, placebo-controlled exploratory study. Aging

clinical and experimental research. Issue, 4, 609-619.

Guralnik, J. M., Simonsick, E. M., Ferrucci, L., Glynn, R. J., Berkman, L. F., Blazer, D. G.,

. . . Wallace, R. B. (1994). A short physical performance battery assessing lower

extremity function: Association with self-reported disability and prediction of mortality

and nursing home admission. Journals of Gerontology, 49(2), M85-M94.

Hamacher, D., Herold, F., Wiegel, P., Hamacher, D., & Schega, L. (2015). Brain activity

during walking: A systematic review. Neuroscience and Biobehavioral Reviews, 57,

310-327.

-143-

Hamacher, D., Singh, N. B., Van Dieen, J. H., Heller, M. O., & Taylor, W. R. (2011).

Kinematic measures for assessing gait stability in elderly individuals: a systematic

review. Journal of the Royal Society Interface, 8(65), 1682-1698.

Hamilton, F., Rochester, L., Paul, L., Rafferty, D., O'Leary, C., & Evans, J. (2009). Walking

and talking: an investigation of cognitive—motor dual tasking in multiple sclerosis.

Multiple Sclerosis Journal, 15(10), 1215-1227.

Hausdorff, J. M. (2005). Gait variability: methods, modeling and meaning. Journal of

NeuroEngineering and Rehabilitation, 2(1), 19.

Hausdorff, J. M. (2007). Gait dynamics, fractals and falls: Finding meaning in the stride-to-

stride fluctuations of human walking. Human Movement Science, 26(4), 555-589.

Hausdorff, J. M., Ashkenazy, Y., Peng, C.-K., Ivanov, P. C., Stanley, H. E., & Goldberger,

A. L. (2001). When human walking becomes random walking: fractal analysis and

modeling of gait rhythm fluctuations. Physica A: Statistical Mechanics and its

Applications, 302(1), 138-147.

Hausdorff, J. M., Balash, J., & Giladi, N. (2003). Effects of Cognitive Challenge on Gait

Variability in Patients with Parkinson’s Disease. Journal of Geriatric Psychiatry and

Neurology, 16(1), 53-58.

Hausdorff, J. M., Edelberg, H. K., Mitchell, S. L., Goldberger, A. L., & Wei, J. Y. (1997).

Increased gait unsteadiness in community-dwelling elderly fallers. Archives of Physical

Medicine and Rehabilitation, 78(3), 278-283.

Hausdorff, J. M., Ladin, Z., & Wei, J. Y. (1995). Footswitch system for measurement of the

temporal parameters of gait. Journal of Biomechanics, 28(3), 347-351.

Hausdorff, J. M., Nelson, M. E., Kaliton, D., Layne, J. E., Bernstein, M. J., Nuernberger, A.,

& Fiatarone Singh, M. A. (2001). Etiology and modification of gait instability in older

-144-

adults: a randomized controlled trial of exercise. Journal of Applied Physiology, 90(6),

2117-2129.

Hausdorff, J. M., Peng, C. K., Ladin, Z., Wei, J. Y., & Goldberger, A. L. (1995). Is walking

a random walk? Evidence for long-range correlations in stride interval of human gait.

Journal of Applied Physiology, 78(1), 349-358.

Hausdorff, J. M., Purdon, P. L., Peng, C. K., Ladin, Z., Wei, J. Y., & Goldberger, A. L.

(1996). Fractal dynamics of human gait: stability of long-range correlations in stride

interval fluctuations. Journal of Applied Physiology, 80(5), 1448-1457.

Hausdorff, J. M., Rios, D. A., & Edelberg, H. K. (2001). Gait variability and fall risk in

community-living older adults: A 1-year prospective study. Archives of Physical

Medicine and Rehabilitation, 82(8), 1050-1056.

Hausdorff, J. M., Schweiger, A., Herman, T., Yogev-Seligmann, G., & Giladi, N. (2008).

Dual-Task Decrements in Gait: Contributing Factors Among Healthy Older Adults.

Journals of Gerontology Series a-Biological Sciences and Medical Sciences, 63(12),

1335-1343.

Hausdorff, J. M., Yogev, G., Springer, S., Simon, E. S., & Giladi, N. (2005). Walking is more

like catching than tapping: gait in the elderly as a complex cognitive task. Experimental

Brain Research, 164(4), 541-548.

Hedges, L. V. (1981). Distribution theory for Glass's estimator of effect size and related

estimators. journal of Educational Statistics, 6(2), 107-128.

Herman, T., Giladi, N., Gurevich, T., & Hausdorff, J. M. (2005). Gait instability and fractal

dynamics of older adults with a “cautious” gait: why do certain older adults walk

fearfully? Gait & Posture, 21(2), 178-185.

-145-

Herssens, N., Verbecque, E., Hallemans, A., Vereeck, L., Van Rompaey, V., & Saeys, W.

(2018). Do spatiotemporal parameters and gait variability differ across the lifespan of

healthy adults? A systematic review. Gait & Posture, 64, 181-190.

Higgins, J., & Green, S. (editors). (2011). Cochrane Handbook for Systematic Reviews of

Interventions, Version 5.1.0 [updated March 2011]. The Cochrane Collaboration.

Retrieved from Available from www.handbook.cochrane.org.

Hsu, Y. L., Chung, P. C., Wang, W. H., Pai, M. C., Wang, C. Y., Lin, C. W., . . . Wang, J. S.

(2014). Gait and balance analysis for patients with Alzheimer's disease using an inertial-

sensor-based wearable instrument. Journal of Biomedical & Health Informatics, 18(6),

1822-1830.

Ijmker, T., & Lamoth, C. J. (2012). Gait and cognition: the relationship between gait stability

and variability with executive function in persons with and without dementia. Gait &

Posture, 35(1), 126-130.

Jansen, I., Heymsfield, S., Baumgartner, R., & Ross, R. (2000). Estimation of skeletal muscle

mass by bioelectrical impedance analysis. Journal of Applied Physiology, 89(2), 465-

471.

Kaufman, A. S., & Lichtenberger, E. O. (1999). Essentials: of WAIS-III assessment. New

York: J. Wiley & Sons.

Kelly, V. E., Eusterbrock, A. J., & Shumway-Cook, A. (2012). A Review of Dual-Task

Walking Deficits in People with Parkinson’s Disease: Motor and Cognitive

Contributions, Mechanisms, and Clinical Implications. Parkinson's Disease, 2012(1), 14.

Kikkert, L., Vuillerme, N., van Campen, J. P., Appels, B. A., Hortobagyi, T., & Lamoth, C.

J. (2017). Gait characteristics and their discriminative power in geriatric patients with

and without cognitive impairment. Journal of NeuroEngineering and Rehabilitation,

14(1), 84.

-146-

Koenig, H. G., Westlund, R. E., George, L. K., Hughes, D. C., Blazer, D. G., & Hybels, C.

(1993). Abbreviating the Duke Social Support Index for use in chronically ill elderly

individuals. Psychosomatics, 34(1), 61-69.

König, A., Klaming, L., Pijl, M., Demeurraux, A., David, R., & Robert, P. (2017). Objective

measurement of gait parameters in healthy and cognitively impaired elderly using the

dual-task paradigm. Aging Clinical & Experimental Research, 29(6), 1181-1189.

Kvelde, T., Lord, S. R., Close, J. C. T., Reppermund, S., Kochan, N. A., Sachdev, P., . . .

Delbaere, K. (2015). Depressive symptoms increase fall risk in older people, independent

of antidepressant use, and reduced executive and physical functioning. Archives of

Gerontology and Geriatrics, 60(1), 190-195.

Lamoth, C. J., van Deudekom, F. J., van Campen, J. P., Appels, B. A., de Vries, O. J., &

Pijnappels, M. (2011). Gait stability and variability measures show effects of impaired

cognition and dual tasking in frail people. Journal of Neuroengineering & Rehabilitation,

8, 2.

Leal, S. L., & Yassa, M. A. (2013). Perturbations of neural circuitry in aging, mild cognitive

impairment, and Alzheimer's disease. Ageing Research Reviews, 12(3), 823-831.

Lee, J., & Park, S. (2018). Effects of a priority-based dual task on gait velocity and variability

in older adults with mild cognitive impairment. Journal of Exercise Rehabilitation, 14(6),

993-997.

Leech, R., Kamourieh, S., Beckmann, C. F., & Sharp, D. J. (2011). Fractionating the Default

Mode Network: Distinct Contributions of the Ventral and Dorsal Posterior Cingulate

Cortex to Cognitive Control. The Journal of Neuroscience, 31(9), 3217-3224.

Leech, R., & Sharp, D. J. (2014). The role of the posterior cingulate cortex in cognition and

disease. Brain, 137(1), 12-32.

-147-

Lehrner, J., Gufler, R., Guttmann, G., Maly, J., Gleiß, A., Auff, E., & Dal-Bianco, P. (2005).

Annual conversion to Alzheimer disease among patients with memory complaints

attending an outpatient memory clinic: the influence of amnestic mild cognitive

impairment and the predictive value of neuropsychological testing. Wiener Klinische

Wochenschrift, 117(18), 629-635.

Lezak, M. D., Howieson, D. B., Loring, D. W., & Fischer, J. S. (2004). Neuropsychological

assessment: Oxford University Press, USA.

Li, K. Z. H., Bherer, L., Mirelman, A., Maidan, I., & Hausdorff, J. M. (2018). Cognitive

Involvement in Balance, Gait and Dual-Tasking in Aging: A Focused Review From a

Neuroscience of Aging Perspective. Frontiers in Neurology, 9(1), 913.

Lin, Y. C., Hsu, W. C., Wu, C. K., Chang, W. H., Wu, K. P. H., & Wong, A. M. K. (2016).

Comparison of motor performance of upper and lower extremities in dual-task tests in

patients with mild Alzheimer's dementia. Aging Clinical and Experimental Research,

28(3), 491-496.

Liu-Ambrose, T. Y., Ashe, M. C., Graf, P., Beattie, B. L., & Khan, K. M. (2008). Increased

risk of falling in older community-dwelling women with mild cognitive impairment.

Physical therapy, 88(12), 1482-1491.

Lord, S., Howe, T., Greenland, J., Simpson, L., & Rochester, L. (2011). Gait variability in

older adults: A structured review of testing protocol and clinimetric properties. Gait &

Posture, 34(4), 443-450.

Lukaski, H., Bolonchuk, W., Hall, C., & Siders, W. (1986). Validation of tetrapolar

bioelectrical impedance method to assess human body composition. Journal of Applied

Physiology, 60(4), 1327-1332.

Lundin-Olsson, L., Nyberg, L., & Gustafson, Y. (1997). “Stops walking when talking” as a

predictor of falls in elderly people. The Lancet, 349(9052), 617.

-148-

Maki, B. E. (1997). Gait Changes in Older Adults: Predictors of Falls or Indicators of Fear?

Journal of the American Geriatrics Society, 45(3), 313-320.

Maquet, D., Lekeu, F., Warzee, E., Gillain, S., Wojtasik, V., Salmon, E., . . . Croisier, J. L.

(2010). Gait analysis in elderly adult patients with mild cognitive impairment and

patients with mild Alzheimer's disease: simple versus dual task: a preliminary report.

Clinical Physiology & Functional Imaging, 30(1), 51-56.

Martin, K. L., Blizzard, L., Wood, A. G., Srikanth, V., Thomson, R., Sanders, L. M., &

Callisaya, M. L. (2013). Cognitive Function, Gait, and Gait Variability in Older People:

A Population-Based Study. The Journals of Gerontology: Series A, 68(6), 726-732.

Martinez-Ramirez, A., Martinikorena, I., Lecumberri, P., Gomez, M., Millor, N., Casas-

Herrero, A., . . . Izquierdo, M. (2016). Dual Task Gait Performance in Frail Individuals

with and without Mild Cognitive Impairment. Dementia & Geriatric Cognitive

Disorders, 42(1-2), 7-16.

Mavros, Y., Gates, N., Wilson, G. C., Jain, N., Meiklejohn, J., Brodaty, H., . . . Fiatarone

Singh, M. A. (2017). Mediation of Cognitive Function Improvements by Strength Gains

After Resistance Training in Older Adults with Mild Cognitive Impairment: Outcomes

of the Study of Mental and Resistance Training. Journal of the American Geriatrics

Society, 65(3), 550-559.

McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., & Stadlan, E. M. (1984).

Clinical diagnosis of Alzheimer's disease Report of the NINCDS ‐ADRDA W o

Group* under the auspices of Department of Health and Human Services Task Force on

Alzheimer's Disease. Neurology, 34(7), 939-939.

Menant, J. C., Schoene, D., Sarofim, M., & Lord, S. R. (2014). Single and dual task tests of

gait speed are equivalent in the prediction of falls in older people: a systematic review

and meta-analysis. Ageing Research Reviews, 16, 83-104.

-149-

Meulen, E., Zijlstra, G., Ambergen, T., & Kempen, G. I. (2014). Effect of Fall ‐Related

Concerns on Physical, Mental, and Social Function in Community ‐Dwelling Ol

Adults: A Prospective Cohort Study. Journal of the American Geriatrics Society, 62(12),

2333-2338.

Meyer, J. S., Xu, G., Thornby, J., Chowdhury, M. H., & Quach, M. (2002). Is Mild Cognitive

Impairment Prodromal for Vascular Dementia Like Alzheimer's Disease? Stroke, 33(8),

1981-1985.

Modarresi, S., Divine, A., Grahn, J. A., Overend, T. J., & Hunter, S. W. (2018). Gait

parameters and characteristics associated with increased risk of falls in people with

dementia: a systematic review. International Psychogeriatrics, 1-17.

Montero-Odasso, M., Bergman, H., Phillips, N. A., Wong, C. H., Sourial, N., & Chertkow,

H. (2009). Dual-tasking and gait in people with mild cognitive impairment. The effect of

working memory. BMC Geriatrics, 9(1), 41.

Montero-Odasso, M., Casas, A., Hansen, K. T., Bilski, P., Gutmanis, I., Wells, J. L., & Borrie,

M. J. (2009). Quantitative gait analysis under dual-task in older people with mild

cognitive impairment: a reliability study. Journal of Neuroengineering & Rehabilitation,

6(1), 35.

Montero-Odasso, M., Muir, S. W., & Speechley, M. (2012). Dual-task complexity affects

gait in people with mild cognitive impairment: the interplay between gait variability, dual

tasking, and risk of falls. Archives of Physical Medicine & Rehabilitation, 93(2), 293-

299.

Montero-Odasso, M., Oteng-Amoako, A., Speechley, M., Gopaul, K., Beauchet, O.,

Annweiler, C., & Muir-Hunter, S. W. (2014). The motor signature of mild cognitive

impairment: results from the gait and brain study. Journals of Gerontology Series A-

Biological Sciences & Medical Sciences, 69(11), 1415-1421.

-150-

Montero-Odasso, M., Verghese, J., Beauchet, O., & Hausdorff, J. M. (2012). Gait and

cognition: A complementary approach to understanding brain function and the risk of

falling. Journal of the American Geriatrics Society, 60(11), 2127-2136.

Moon, Y., Sung, J., An, R., Hernandez, M. E., & Sosnoff, J. J. (2016). Gait variability in

people with neurological disorders: A systematic review and meta-analysis. Human

Movement Science, 47, 197-208.

Muir-Hunter, S. W., & Wittwer, J. E. (2016). Dual-task testing to predict falls in community-

dwelling older adults: a systematic review. Physiotherapy, 102(1), 29-40.

Muir, S. W., Gopaul, K., & Montero Odasso, M. M. (2012). The role of cognitive impairment

in fall risk among older adults: a systematic review and meta-analysis. Age and ageing,

41(3), 299-308.

Muir, S. W., Speechley, M., Wells, J., Borrie, M., Gopaul, K., & Montero-Odasso, M. (2012).

Gait assessment in mild cognitive impairment and Alzheimer's disease: The effect of

dual-task challenges across the cognitive spectrum. Gait and Posture, 35(1), 96-100.

Nakagawa, S. (2004). A farewell to Bonferroni: the problems of low statistical power and

publication bias. Behavioral Ecology, 15(6), 1044-1045.

Nascimbeni, A., Caruso, S., Salatino, A., Carenza, M., Rigano, M., Raviolo, A., & Ricci, R.

(2015). Dual task-related gait changes in patients with mild cognitive impairment.

Functional Neurology, 30(1), 59-65.

Nasreddine, Z. S., Phillips, N. A., Bédirian, V., Charbonneau, S., Whitehead, V., Collin, I., .

. . Chertkow, H. (2005). The Montreal Cognitive Assessment, MoCA: A Brief Screening

Tool For Mild Cognitive Impairment. Journal of the American Geriatrics Society, 53(4),

695-699.

-151-

Ng, F., Trauer, T., Dodd, S., Callaly, T., Campbell, S., & Berk, M. (2007). The validity of

the 21-item version of the Depression Anxiety Stress Scales as a routine clinical outcome

measure. Acta Neuropsychiatrica, 19(5), 304-310.

Painter, J. A., Allison, L., Dhingra, P., Daughtery, J., Cogdill, K., & Trujillo, L. G. (2012).

Fear of Falling and Its Relationship With Anxiety, Depression, and Activity Engagement

Among Community-Dwelling Older Adults. American Journal of Occupational Therapy,

66(2), 169-176.

Petersen, R. C. (1999). Mild Cognitive Impairment: Clinical Characterization and Outcome.

The Journal of the American Medical Association, 281(19), 1777.

Petersen, R. C., Doody, R., Kurz, A., & et al. (2001). Current concepts in mild cognitive

impairment. Archives of Neurology, 58(12), 1985-1992.

Pettersson, A., Olsson, E., & Wahlund, L.-O. (2005). Motor function in subjects with mild

cognitive impairment and early Alzheimer’s disease. Dementia and Geriatric Cognitive

Disorders, 19(5-6), 299-304.

Pirker, W., & Katzenschlager, R. (2017). Gait disorders in adults and the elderly. Wiener

klinische Wochenschrift, 129(3), 81-95.

Plummer, P., Zukowski, L. A., Giuliani, C., Hall, A. M., & Zurakowski, D. (2015). Effects

of Physical Exercise Interventions on Gait-Related Dual-Task Interference in Older

Adults: A Systematic Review and Meta-Analysis. Gerontology, 62(1), 94-117.

Pointer, S. C. (2018). Trends in hospitalised injury, Australia 1999-00 to 2014-15 Injury

research and statistics series no. 110 (Vol. Cat. no. INJCAT 190). Canberra: AIHW.

Rabin, L. A., Paré, N., Saykin, A. J., Brown, M. J., Wishart, H. A., Flashman, L. A., &

Santulli, R. B. (2009). Differential memory test sensitivity for diagnosing amnestic mild

cognitive impairment and predicting conversion to Alzheimer's disease.

-152-

Neuropsychology, development, and cognition. Section B, Aging, neuropsychology and

cognition, 16(3), 357-376.

Rhea, C. K., Silver, T. A., Hong, S. L., Ryu, J. H., Studenka, B. E., Hughes, C. M. L., &

Haddad, J. M. (2011). Noise and Complexity in Human Postural Control: Interpreting

the Different Estimations of Entropy. PLoS ONE [Electronic Resource], 6(3), e17696.

Rosano, C., Aizenstein, H., Brach, J., Longenberger, A., Studenski, S., & Newman, A. B.

(2008). Special ArticleGait Measures Indicate Underlying Focal Gray Matter Atrophy in

the Brain of Older Adults. The Journals of Gerontology: Series A, 63(12), 1380-1388.

Rossiter-Fornoff, J. E., Wolf, S. L., Wolfson, L. I., & Buchner, D. M. (1995). A cross-

sectional validation study of the FICSIT common data base static balance measures.

Frailty and Injuries: Cooperative Studies of Intervention Techniques. The journals of

gerontology. Series A, Biological sciences and medical sciences, 50(6), M291.

Royall, D. R., Lauterbach, E. C., Cummings, J. L., Reeve, A., Rummans, T. A., Kaufer, D.

I., . . . Coffey, C. E. (2002). Executive Control Function. The Journal of Neuropsychiatry

and Clinical Neurosciences, 14(4), 377-405.

Rubenstein, L. Z. (2006). Falls in older people: epidemiology, risk factors and strategies for

prevention. Age and ageing, 35(suppl_2), ii37-ii41.

Rucco, R., Agosti, V., Jacini, F., Sorrentino, P., Varriale, P., De Stefano, M., . . . Sorrentino,

G. (2017). Spatio-temporal and kinematic gait analysis in patients with Frontotemporal

dementia and Alzheimer's disease through 3D motion capture. Gait & Posture, 52, 312-

317.

Ryff, C. D., & Keyes, C. L. M. (1995). The Structure of Psychological Well-Being Revisited.

Journal of Personality and Social Psychology, 69(4), 719-727. doi: 10.1037/0022-

3514.69.4.719

-153-

Schwenk, M., Zieschang, T., Englert, S., Grewal, G., Najafi, B., & Hauer, K. (2014).

Improvements in gait characteristics after intensive resistance and functional training in

people with dementia: a randomised controlled trial. BMC Geriatrics, 14(1), 73.

Shaw, F. E. (2002). Falls in cognitive impairment and dementia. Clinics in Geriatric

Medicine, 18(2), 159-173.

Sheline, Y. I., Wang, P. W., Gado, M. H., Csernansky, J. G., & Vannier, M. W. (1996).

Hippocampal atrophy in recurrent major depression. Proceedings of the National

Academy of Sciences, 93(9), 3908-3913.

Sheridan, P. L., Solomont, J., Kowall, N., & Hausdorff, J. M. (2003). Influence of executive

function on locomotor function: divided attention increases gait variability in

Alzheimer's disease. Journal of the American Geriatrics Society, 51(11), 1633-1637.

Smith, E., Cusack, T., Cunningham, C., & Blake, C. (2017). The Influence of a Cognitive

Dual Task on the Gait Parameters of Healthy Older Adults: A Systematic Review and

Meta-Analysis. Journal of Aging & Physical Activity, 25(4), 671-686.

Snijders, A. H., van de Warrenburg, B. P., Giladi, N., & Bloem, B. R. (2007). Neurological

gait disorders in elderly people: clinical approach and classification. The Lancet

Neurology, 6(1), 63-74.

Springer, S., Giladi, N., Peretz, C., Yogev, G., Simon, E. S., & Hausdorff, J. M. (2006). Dual-

tasking effects on gait variability: The role of aging, falls, and executive function.

Movement Disorders, 21(7), 950-957.

St Clair-Thompson, H. L., & Allen, R. J. (2013). Are forward and backward recall the same?

A dual-task study of digit recall. Memory & Cognition, 41(4), 519-532.

Stergiou, N., & Decker, L. M. (2011). Human movement variability, nonlinear dynamics, and

pathology: Is there a connection? Human Movement Science, 30(5), 869-888.

-154-

Stewart Williams, J., Kowal, P., Hestekin, H., O’Driscoll, T., Peltzer, K., Yawson, A., . . .

collaborators, S. (2015). Prevalence, risk factors and disability associated with fall-

related injury in older adults in low- and middle-incomecountries: results from the WHO

Study on global AGEing and adult health (SAGE). BMC Medicine, 13(1), 147.

Suo, C., Fiatarone Singh, M., Gates, N., Wen, W., Sachdev, P., Brodaty, H., . . . Singh, N.

(2016). Therapeutically relevant structural and functional mechanisms triggered by

physical and cognitive exercise. Molecular Psychiatry, 21(11), 1633.

Taylor, M. E., Delbaere, K., Mikolaizak, A. S., Lord, S. R., & Close, J. C. (2013). Gait

parameter risk factors for falls under simple and dual task conditions in cognitively

impaired older people. Gait & Posture, 37(1), 126-130.

Tian, Q., Chastan, N., Bair, W.-N., Resnick, S. M., Ferrucci, L., & Studenski, S. A. (2017).

The brain map of gait variability in aging, cognitive impairment and dementia—a

systematic review. Neuroscience & Biobehavioral Reviews, 74, 149-162.

van Iersel, M. B., Hoefsloot, W., Munneke, M., Bloem, B. R., & Olde Rikkert, M. G. M.

(2004). Systematic review of quantitativeclinical gait analysis in patients with dementia.

Zeitschrift für Gerontologie und Geriatrie, 37(1), 27-32.

Verghese, J., LeValley, A., Hall, C. B., Katz, M. J., Ambrose, A. F., & Lipton, R. B. (2006).

Epidemiology of Gait Disorders in Community-Residing Older Adults. Journal of the

American Geriatrics Society, 54(2), 255-261.

Verhaeghen, P., Steitz, D. W., Silwinski, M. J., & Cerella, J. (2003). Aging and Dual-Task

Performance: A Meta-Analysis. Psychology and Aging, 18(3), 443-460.

Wajda, D. A., Mirelman, A., Hausdorff, J. M., & Sosnoff, J. J. (2017). Intervention modalities

for targeting cognitive-motor interference in individuals with neurodegenerative disease:

a systematic review. Expert Review of Neurotherapeutics, 17(3), 251-261.

-155-

Whooley, M. A., Kip, K. E., Cauley, J. A., Ensrud, K. E., Nevitt, M. C., Browner, W. S., &

Group, f. t. S. o. O. F. R. (1999). Depression, Falls, and Risk of Fracture in Older Women.

Archives of Internal Medicine, 159(5), 484-490.

Winblad, B., Palmer, K., Kivipelto, M., Jelic, V., Fratiglioni, L., Wahlund, L.-O., . . .

Petersen, R. C. (2004). Mild cognitive impairment – beyond controversies, towards a

consensus: report of the International Working Group on Mild Cognitive Impairment.

Journal of Internal Medicine, 256(3), 240-246.

Wittwer, J. E., Webster, K. E., & Hill, K. (2014). The effects of a concurrent motor task on

walking in Alzheimer's disease. Gait & Posture, 39(1), 291-296.

Yesavage, J. A., Brink, T. L., Rose, T. L., Lum, O., Huang, V., Adey, M., & Leirer, V. O.

(1982). Development and validation of a geriatric depression screening scale: a

preliminary report. Journal of psychiatric research, 17(1), 37-49.

Yogev-Seligmann, G., Hausdorff, J. M., & Giladi, N. (2008). The role of executive function

and attention in gait. Movement Disorders, 23(3), 329-342.

Zhang, W., Low, L.-F., Gwynn, J. D., & Clemson, L. (2019). Interventions to Improve Gait

in Older Adults with Cognitive Impairment: A Systematic Review. Journal of the

American Geriatrics Society, 67(2), 381-391.

Zimmerman, M. E., Lipton, R. B., Pan, J. W., Hetherington, H. P., & Verghese, J. (2009).

MRI- and MRS-Derived Hippocampal Correlates of Quantitative Locomotor Function

in Older Adults. Brain research, 1291, 73-81.